DRAFT Minutes: Credible Web CG (26-27 July 2018)



In person (for one or more session):

  1. An Xiao Mina (co-chair, Meedan and Credibility Coalition)
  2. Annette Greiner (LBL)
  3. Aviv Ovadya - day 1 afternoon and dinner
  4. Ben Werdmuller (Matter Ventures) - day 2 only
  5. Chris Guess (Duke Reporters Lab)
  6. Dan Brickley (Google -
  7. Ed Bice (Meedan)
  8. Jon Udell (
  9. Nick Pickles (Twitter) - day 1 morning
  10. Pat Hayes
  11. Sandro Hawke (co-chair, W3C/MIT, Hawkeworks)
  12. Scott Lowenstein (Twitter)
  13. Scott Yates (CableLabs - Certified Content Coalition)
  14. Tantek Çelik (Mozilla)
  15. Tom Gilbert - day 1 only
  16. Vinny Green (Snopes)
  17. Zoë Triska (Facebook

Remote (for one or more session): (Note that room acoustics were difficult and made parts of the meeting unintelligible)

  1. Cheryl Langdon-Orr (Internet Australia/ICANN-ALAC/At-Large)
  2. Davide Ceolin (CWI Amsterdam)
  3. Ed Summers (
  4. Farnaz Jahanbakhsh (MIT)
  5. Greg McVerry (Southern Connecticut State University)
  6. Jesse Kranzler (trustmetrics)
  7. Michael Golebiewski (Microsoft)
  8. Newton Calegari (
  9. Olaf Steenfadt (Reporters Without Borders)
  10. Reto Gmür (FactsMission)
  11. Sally Lehrman (The Trust Project)
  12. Tzviya Siegman (Wiley)
  13. Vagner Diniz (


@@todo – anyone reading through this, please suggest where topic headings should go



Minutes Day 1

# [Sandro Hawke] *Face to face meeting!*

# [Cheryl Langdon-Orr] I was getting lonely here ;-)

# [Reto Gmür] present +

# [An Xiao Mina] attendee list:

# [Sandro Hawke] introductions

# [Dan Brickley] Sandro and An introduce themselves, Sandro notes Facebook and Google funding to support his work at W3C on this (Not scribing all intros)

# [Dan Brickley] q for remote people, how is the audio?

# [Cheryl Langdon-Orr] answer re audio - variable but OK

# [Michael Golebiewski] Decent, but fading in an out

# [Annette Greiner] I’m from Lawrence Berkeley National Lab, working as a web app developer. I have a long-standing interest in science communication, stemming in part from many years’ work in the past as a science writer and editor. Before that, I did biomedical research. Now I work with researchers to help them share data.

# [Tantek] Tantek Çelik, Mozilla

# [Dan Brickley] can we all type a sentence or two intro for the meeting records?

# [Cheryl Langdon-Orr] I linked my name in that list to my LinkedIn profile … Or just Google me ;-)… But think of my perspectives as that of “the Internet end user” or the ‘consumer/public interest’…

# [Dan Brickley] mine is that I’m Dan Brickley, I work at Google in SF, originally worked on trust/quality credibility stuff in the ’90s “subject based catalogues” community (which failed!), more recently running project, and engaging with the fact checking community via the ClaimReview collaboration.

# [Zoë Triska] I’m from Facebook, specifically the Facebook Journalism Project. I focus on elevating credible and reliable news sources on the platform. Before that, I was at HuffPost, working on product and editorial strategy for their international editions.

# [Tzviya Siegman] I work for Wiley, a scholarly publisher. I chair the W3C Publishing WG and am on the W3C Advisory Board. I also have an interest in Verifiable Claims and methods of fact checking and proving things to be “true”.

# [Tom Gilbert] I’m a Ph.D. candidate at Berkeley in Machine Ethics and Epistemology, also a member of a Berkeley-affiliated Goodly Labs project called Public Editor, that works to combat misinformation by confronting cognitive biases.

# [Sandro Hawke] next: Ed, Farnaz, Cyndi, Cheryl, Davida, Greg, Michael

# [Vinny Green] I’m the Vice President of Operations for One of my many responsibilities is to lead the development projects for including all the scalable fact checking efforts. Prior to, I co-founded an advertising technology company and previously served in the U.S. Marine Corps.

# [Dan Brickley] (I’m having trouble hearing the remote audio, Reto’s at least; hmm Edsu’s is a lot clearer it seems)

# [Scott Lowenstein] I’m working at Twitter for the summer on better understanding identity and credibility after spending many years in politics and political campaigns.

# [An Xiao Mina] I’m An, co-chair of this group and co-founder of the Credibility Coalition, where we’re building standards for assessing content credibility online through rigorous research and open data: Director of Product at Meedan, where we build tools for global journalism and translation. Research affiliate at the Berkman Klein Center at Harvard.

# [Dan Brickley] (Greg can’t get onto Zulip, according to a msg that popped up on screen here briefly)

Farnaz: I’m a PhD student at MIT CSAIL, research designing and building tools to help people assess accuracy of news and credibility of news sources according to their criteria

# [Davide Ceolin] I’m a tenure-track researcher at CWI Amsterdam, focusing on information quality and trust online. Working on combining crowdsourcing and nichesourcing with automated reasoning.

# [Cheryl Langdon-Orr] my audio has faded

# [Ed Summers] I’m Ed Summers a researcher at the University of Maryland where I’m studying the role that web archives play in online discourse. I’m particularly interested in the role that web archives like the Internet Archive play in determining the provenance of web resources, and what, if anything, they have to do with credibility on the web.

# [Michael Golebiewski] The room audio is gone

# [Dan Brickley] (discussion of audio quality)

# [Cheryl Langdon-Orr] nope

# [Cheryl Langdon-Orr] table mic isn’t picking you up now (even the loud chair scraping has dulled

# [Scott Yates] I’m Scott Yates, founder of the Certified Content Coalition, a new initiative from CabelLabs, which is the research arm of the cable industry. The CCC is a new certification authority that will be able certify to basic journalism standards like transparency and editorial process. My background is in journalism and three different startups that I started and exited.

# [Dan Brickley] Sandro: “If you can tell I’m talking, I don’t think there’s anything else we can do … but try to project”

# [Dan Brickley] Michael: I work for Microsoft on Bing, trustworthy AI and computing, …

# [Dan Brickley] (is it useful if I scribe more literally, given the audio?)

# [Dan Brickley] Sandro: next 3 things I wanted to do in the getting started, … and ok if it leaks into next section, …

# [Reto Gmür] but the sound was much louder before

# … and in general the work of the group(s)

# … most of you should be familiar by now, the breakdown into 4 subgroups, plus a 5th one

# … I wrote them on the wall

# [Dan Brickley] 1 - Inspection 2- Corroboration

# [Dan Brickley] 3- Reputation

# [Dan Brickley] 4- Transparency

# [Dan Brickley] and the new 5 - Social Reaction

# [Dan Brickley] Sandro: these are aspects of figuring out if something is true

# … 1st is most aligned with the credibility coalition’s work

# [Dan Brickley] (discussion of why the remote video disspeared here)

# [Ed Summers] yeah – the only thing i can hear well is laughter, lol

# [Dan Brickley] (Chris wrestles heroically with the missing video)

# [Dan Brickley] (Chris tries the old turn-it-off-and-on again thing with the cable)

# [Dan Brickley] (video restored)

# [Dan Brickley] Sandro: Area 2, you’re looking at the thing, trying to understand if it’s true, … so you look elsewhere. Factchecking is the main activity here, but also things like wikipedia

# … AREA 3, reputation eg. citation patterns

# [Dan Brickley] Area 4: transparency, sites saying more about themselves, … some vulnerability to lying, but a piece of the puzzle

# [Dan Brickley] plus new Area 5: social reaction, … you look around to see what other people are doing/saying

# [Dan Brickley] e.g. at Facebook when things go viral, they can route them to factcheckers for inspection/review

# [Dan Brickley] Sandro: On the agenda, we basically take turns through these … arrangement/order is based on who could join when. E.g. transparency we have Sally from Trust Project and Olaf from […missed] calling in. Any Qs re agenda?

# [Dan Brickley] How is remote audio?

# [Dan Brickley] Ed: we’re looking for a microphone

# [Dan Brickley] (audio engineering discussions)

# [Cheryl Langdon-Orr] Audio from the room is still CRAP

# [Dan Brickley] (Sandro tries the old leave-the-meeting-and-rejoin trick)

# [Dan Brickley] (Sandro hits the end-meeting-for-all button by accident)

# [Cheryl Langdon-Orr] ohh dear that explains a lot…

# [Dan Brickley] (democracy, peace and justice splutter back into life, thanks to our efforts)

# [Dan Brickley] /me misses Zakim

# [Dan Brickley] (background noise in the room from other nearby groups rises omminously)

# [Dan Brickley] Sandro: how is audio?

# [Dan Brickley] (TV says, Chrome OS: “Mirroring to Panasonic-TV”)

# [Dan Brickley] (rejoining…)

# [Cheryl Langdon-Orr] *and the audio is still poor *yet it was OK at the start upof the meeting**

# [Tantek] (video conferencing is hard, projecting is hard, doing both together … ?)

# [Tzviya Siegman] (@Dan Brickley don’t we all miss Zakim? I am on IRC if you wish to join me. :))

# [Tantek] (/me checks IRC)

# [Dan Brickley] Sandro: can anyone remote hear me?

# [Ed Summers] audio seems a bit better now, from sandro

# [Cheryl Langdon-Orr] we can hear you @Sandro Hawke

# [Dan Brickley] Tzviya: It’s suddenly working again!

# [Dan Brickley] /me dials back on the literal scribing; let me know if more needed

# [Cheryl Langdon-Orr] Thanks @Dan Brickley apprecited

# [Dan Brickley] (can we pin “” in the UI here somewhere?)

# [Dan Brickley] Vagner (from Brazil), I work for, org responsible for domain names within .br, … many issues. I’m responsible for the Web & Technology studies centre here

# … issues around credibility

# [Sandro Hawke] I go to then click on Agenda…

# … there will be an election in November here

# … we expect to be a target for fake news, all those things related to news/elections, … hope to learn

# [Dan Brickley] Jessie: At Trustmetrics in NYC. We primarily in brand safety and quality space. Last couple months, shifting towards fake news and dissinformation

# … interested to aggregate lots of signals from diff players throughout the industry

# … help platforms and others make … decisions

# [Dan Brickley] (collective admiration of the high quality Jessie’s audio quality)

# [Dan Brickley] Sandro: Next topic, … survey of related work in this space and how to understand/organize it

# … couple of different efforts lately that are starting to coallesce

# [Scott Yates] Landscape

# … we could look at that, building on recent telecon. Starting with [one of the] Scott(s)

# [Dan Brickley] ScottY: [shared url above] certified content coalition. Initially surveyed for efforts we could join, standards based bodies and existing orgs

# … I started this ridiculously large list of all the efforts that are out there, updated today,… see 2nd tab with list-of-lists for similar efforts that have informed mine

# … just alphabetical and rough guess at org type. SOmetimes a blend of corp-gov-thinktank themes, category of what the solution is. For example, ICFQ, a great nonprofit org, … but their core mission is not anything to do with misinformation

# … but they have a journalstic standards activity

# … This is editable, feel free to add other efforts into this, I’d appreciate if you let me know if you do this, … as I keep other things in sync. Happy to discuss / collaborate

# [Dan Brickley] Sandro: [discusses spreadsheet structure] …

# [Dan Brickley] Dan: can people add columns without disussion?

# [Dan Brickley] ScottY: happy with new columns!

# [An Xiao Mina]

# [Dan Brickley] An: I’m going to put a list in that we worked on at CUNI a few weeks ago, …

# …misinfomaps

# [Dan Brickley] –^^

# … a lot of lists that people are developing

# … understand the landscape in terms of categories, an x-axis left to right, from theory to action, vertical going from content to infrastructure/algos

# …. meant to be additive. For CredCo somewhere in the middle

# … very bare bones framework

# … but it was helpful for us, especially on the action side of things

# [Dan Brickley] ScottY: That’s great, just added it to my list of lists

# [Dan Brickley] An: … we could visualize the [spreadsheet items?] into this map

# [Dan Brickley] Vinny: someone who knows the Google sheets API could do something useful here maybe

# [Dan Brickley] An: We’re joined by my colleague Micheala (sp?) from Medan

# [Dan Brickley] Sandro: now we break into 3-ish people groups

# … pick a couple bits on this list

# [Dan Brickley] if you’re group 1, … put that on the spreadsheet, … and try to see where it fits on the content-vs-theory (C vs T %) axis, and the theory-to-action % axis

# … but mostly which quadrant it’s in

# [Dan Brickley] Sando: may also want to consider where it is geographically, country/continent

# … looking for which ones are seeking volunteers

# [Dan Brickley] ScottY: we can make columns E-F numeric, making them rankable

# [Dan Brickley] Sandro: I feel weird about negative numbers

# [Dan Brickley] /me ..ooOO( complex numbers? )

# [Dan Brickley] ScottY: any recommendations?

# [Dan Brickley] ScottY: Ok, if all content is -10, all mechanism is +10

# [Dan Brickley] Sandro: [missed detail]

# [Dan Brickley] Jon: Many of us won’t know the orgs and won’t have time to do much realtime research

# [Dan Brickley] ScottY/Sandro: couple of minutes to get a rough idea

# [Dan Brickley] content is +10

# [Dan Brickley] Sandro: Intuition is that infrastructure is kind of foundational [hence lower and negative numbers]

# [Dan Brickley] Vinny: [shows screen with project management viz tool]

# … if there’s a library to connect to a data source, … if we knew the name of the visualization style

# [Dan Brickley] (@Vinny - it might be in somewhere…)

# [Annette Greiner]

# [Dan Brickley] thanks Annette!

# [Dan Brickley] (Sandro adding example in spreadsheet)

# [Dan Brickley] (for ClaimBuster, row 15 in )

# [Dan Brickley] (Climate Feedback as 2nd example)

# [Dan Brickley] Claimbuster (10,0); Climate feedback (10,10)

# [Dan Brickley] danbri: suggests Snopes

# [Dan Brickley] Sandro: Snopes seems entirely content, … less infra; and seems entirely practice, … so (10,10); geographically, English language only, but global reach topically

# [Dan Brickley] Vinny: 100% focussed on English speakers, predom US, but global audience

# [Dan Brickley] (I’m stopping scribing for a while)

# [Cheryl Langdon-Orr] OK so now the room BG noise is going to be very “distracting”…

# [Cheryl Langdon-Orr] We can but try @Sandro Hawke

# [Sandro Hawke] Everyone remote - please treat yourself as a subgroup

# [Sandro Hawke] remote = group 5

# [Sandro Hawke] I was going to split you into multiple subgroups, but zoom isn’t giving me that option today :-(

# [Ed Summers] but it’s a nice time to make some ☕️

# [Sandro Hawke] yeah, let me grab my phone

# [Cheryl Langdon-Orr] very true @Ed Summers

# [Dan Brickley] the task is to pick orgs that are rows in the spreadsheet, and fill out their x/y axis values and other fields

# [An Xiao Mina] We discussed here that another way to think about the y axis is the stack - where on the stack is the initiative operating?

# [Vinny Green] Any thoughts with using conditional formatting on the spreadsheet to highlight rows that are complete?

# [An Xiao Mina] *Re: “” – quote Any thoughts with using conditional formatting on the spreadsheet to highlight rows that are complete? *

# [Sandro Hawke] BREAK

# [Dan Brickley] for spreadsheet, something like might help going from the x and y values to 4 quadrant codes

# [Scott Yates] taking over as your new humble scribe

# [Dan Brickley] can some remote people try to talk to check the audio please

# [Cheryl Langdon-Orr] Yes audio good and now vision better framed :-)

# [Scott Yates] fixing video so that remote viewers don’t have to look at a table full of cables

# [Scott Yates] is that view better?

# [Cheryl Langdon-Orr] ok thx

# [Tantek] My favorite example of that (appeal to authority) was the Bill Joy article on nanotech in WIRED.

# [Sandro Hawke] scott, you can just start a line with “..” if it’s the same speaker

# [Scott Yates] ”..” Slide about Scientific Consensus

# .. science is always the best guess

# … Slide about Modes of ensuring integrity

# … Slide about Verification

# … Slide about Education

# …. Slide about The Scientific Method

# … Slide about Documentation

# [Tantek] “Occam’s razor - simplest solution should be the one that you consider, all other things being equal.”

# [Dan Brickley] Accreditation – can we go deeper into that topic at some point please?

# [Dan Brickley] related work,

# …. Slide on Controls

# … slide on How Can Science Foster Trust?

# … slide on Reaching outside the ivory tower

# [Sandro Hawke] Annette’s handout:

# … How can we do something useful using the manners of policing?

# Sandro Hawke says: What we want to do is move from the general to the specific of what could be done technologically in our space

# … What are some tangible next steps?

# … Some computer system could be a help?

# [Sandro Hawke] Later half of alphabet go to

# [Cheryl Langdon-Orr] OK that helped last time :-)

# [Reto Gmür] what happened to audio

# [Cheryl Langdon-Orr] terrible from the room we seem ok with each other

# [Tantek] Regarding “humanizing scientists”, is the counter on-topic? Like how to fight the mocking and sometimes de-humanizing of scientists? e.g. examples of a scientists in pop media

# Reto Gmür says: he is a bit scared because of the troubles involved in “selling” science

# Sandro Hawke says: people trust science based on wrong signals, such as skin color

# [Reto Gmür] selling with non scientif reason like “excited scientist” and “compelling story”

# Christopher Guess says: most people don’t really know about internal reviews.

# … We have to meet people at their level.

# [Michael Golebiewski] I will have to drop off for a brief meeting over lunch. Likely I will miss the breakout and will try to join back for the readouts.

# Ed Bice says: concern is that indicators can and will be gamed

# [Tantek] Or even just getting pop scientific articles updated with citations of retractions for whatever scientific papers they originally cited?

# Christopher Guess says: retractions come out months later

# … for regular people, though, they will be a sidenote at best

# [Tantek] If anything over-emphasis on retractions could just feed reputation smearing of science in general.

# [Sandro Hawke] calling in to talk to remote folks

# [Scott Yates] OK, scribing is done while breakouts start

# [Tzviya Siegman]

# [Scott Yates] Slowly coming back from breakouts….

# [Cheryl Langdon-Orr] indeed!

# [Cheryl Langdon-Orr] :-)

# An Xiao Mina says: from Group 1: Need to recognize difference in audiences.

# … Documentation is difficult to publish, could there be standardization of documentation for publishing

# … could there be standards of reporting of authority of sources.

# Jon Udell says: Group TWO (not 3) discussed the financial motivation of researchers.

# … good discussion, but no clear answer as to how to report

# Nick Pickles says: there could be better work on showing previous articles on a topic from a publication could be more useful than showing related articles from other places.

# … general answer is no

# Nick Pickles says: showing evolution from a publication could be useful

# Vinny Green says: he doesn’t think about the audience for snopes as individuals, but to try to improve “herd immunity.”

# … adds that some research shows fact checking doesn’t work, but info from friends works, and where do friends get their info

# … says group looked especially at the individual who is doing the work

# [Scott Yates] (We’ll have a link to a slide in a moment)

# [Reto Gmür] /me is happy that sound is so much better now, but its getting too late, see you tomoorw

# … so how do you build, for instance, a twitter list that could show some credibility

# Zoë Triska says: conflict between safety and transparency is very tricky

# [Scott Yates] Group 4:

# [Sandro Hawke]

# [Dan Brickley] mention of

# Tantek says: they talked about the example of

# … what would be a way to do an automatic evaluation of content?

# … not aiming at those who ignore fact checking

# [Dan Brickley]

# … could we look for a source of negative templates

# … in the way a browser checks for a valid cert, could there be a system that shows a site has bad science

# [Vinny Green]

# Sandro Hawke says: time to move on

# [Vinny Green] *Group 3 Notes:

# [Tantek] method is to check for existence of negative templates or categories

# Sandro Hawke says: , problem of claim revue on URLs has the issue of AMPs and other FB page versions of stories

# [Dan Brickley] *backstory for Vinny’s notes

Example usecases (that fed into group 3 look at edu/occupational credentials)

Imagine attempting a #factcheckplz tag, where people on Twitter, Pinboard etc could flag URLs for the attention of fact checking organizations such as Snopes It would probably quickly be used agressively and become tricky Are there ways that e.g. Snopes could use better metadata to filter down from that chaos into something useful. For example, people who we have evidence are such-and-so qualified (e.g. went to medical school, have phd in xyz) have flagged something… Those profiles exist, … they speak in a kind of templated mixed voice of the researcher and some scholarly institution, e.g. Technologies (oauth, openidconnect, relmeaauth…) exist that allow people to show control of pages, and reprocated links … Privacy eg consider … firehose of hate being pointed at a particular researcher, discusion of pseudonyms. How could a community build e.g. systems where you could verify say a Twitter list as all having verifiable qualifications in e.g. nuclear power station construction*

# [Cheryl Langdon-Orr] thanks for my morning coffee read @Vinny Green

# [Cheryl Langdon-Orr] Ping us here when you reconvene please as we remote people will need to get back into Zoom

# [Cheryl Langdon-Orr] Joined anyway

# [Cheryl Langdon-Orr] yes audio is OK

# [Sandro Hawke] About to restart

# [Cheryl Langdon-Orr] yes

# [Dan Brickley] (to them, we’re the remote ones…)

# [Cheryl Langdon-Orr] we can hear

# [Sandro Hawke] But can we hear you?

# [Sandro Hawke] Also, I don’t even see you on….. Weird.

# [Sandro Hawke] I see the three room cameras and Vagner

# [Cheryl Langdon-Orr] Wierd indeed Did you just hear me?

# [Sandro Hawke] no, try again….

# [Cheryl Langdon-Orr] we are hearing the room

# [Cheryl Langdon-Orr] I give up!

# [Vinny Green] *Regarding what I am talking about:

@Ed Bice here you go!*

# [Cheryl Langdon-Orr] Well I don’t actually but you get my drift remote participants ;-)

# [Cheryl Langdon-Orr] Although I am in 2 different meetings so perhaps the Adobe Connect Room has stolen my Mike even though I am on the phone to it as well…

# [Sandro Hawke] discussion of size of fact checking community

# [Sandro Hawke] not a lot of tech resources

# Dan Brickley says: biggest problem is whack-a-mole problem, echo problem, where the same claim appears in many forms around the web all the time

# .. How do we scale up the ability to connect Expensive fact checks to these appearnaces

# .. We’ve added a type called Claim, but really it’s a fact check and it’s appearances.

# Christopher Guess says: our experioence at Share The Facts

# .. start with postives

# .. it does a really job of capturing what the journalists want to capture

# .. it covers most of the important bases

# .. which is great. In the JSON format you can show it, and it makes sense

# .. except sameAs field

# [Dan Brickley] (re sameAs, there was an ad-hoc hack that overloaded it with “url where claim appears”)

# .. folks all want their one little extra thing to be added

# .. so sameAs got glommed on and … didnt quite fit right

# Dan Brickley says: there was a particular modeling mistake there, yeah

# [Pat Hayes] foo

# [Sandro Hawke] hey pat.

# [Sandro Hawke] can you scribe here?

# [Sandro Hawke] Chris show a polticfact ShareTheFacts widget

# [Pat Hayes] I can try. I dont know any names :-)

# .. work around CMS limitations

# [Sandro Hawke] I’ll try to call people by name?

# [Vinny Green] What’s the zoom link?

# [Sandro Hawke] chris says technologies backlogged

# [Sandro Hawke]

# [Pat Hayes] its delicate and hard to embed.

# [Sandro Hawke] chris says most folks use the form in 20 seconds now

# [Pat Hayes] GUI very quick average 20-30 secs to use.

# [Pat Hayes] But issue is where its in the pipeline.

# [Dan Brickley] *Sharing in case we discuss later: Bill Adair gave us some great examples ( where we have the “same” reappearing, in this case from the same person:

”"”How about this:

Mitt Romney:

Says Barack Obama began his presidency “with an apology tour.”

Several times, Obama “has apologized for what he deems to be American arrogance, dismissiveness, and derision; for dictating solutions, for acting unilaterally … .”

President Obama “went around the world and apologized for America.””””*

# [Pat Hayes] values for location include things like ‘internet’. Actual URL rare.

# [Dan Brickley] Q+ to ask both Chris and Vinny what consensus they see on definitions for what it is in practice to be the “same claim”. Can anything be written down shared across the wider community, or does it vary site-by-site?

# [Pat Hayes] Discussion of what counts as valid/best/useful fillers for slots like location, author.

# [Pat Hayes] danB: strong sense of vibrancy of organizations doing things but all differently. Should we be looking for common ideas?

# [Pat Hayes] Chris, Vinny: no

# [Pat Hayes] DanB: What are criteria of identification for claims?

# [Pat Hayes] Vinny: every report is unique claim, all else is editorializing.

# [Dan Brickley] vinny: our cms doesn’t yet, but headed that way, … a multi-claim fact check

# [Pat Hayes] Chris: need way to define overarching claim

# [Pat Hayes] Vinny: already done that.

# [Pat Hayes] Ed: be careful about context.

# [Dan Brickley] using “advanced custom fields’ in Wordpress currently, moving towards their newer Gutenberg block system

# [Dan Brickley] can make custom blocks, e.g. redistrib on git(hub etc)

# [Pat Hayes] Vinny: we use data and schema primarily internally, hope it will work for public exposure.

# [Dan Brickley] Vinny: “this widget… would be if the same org made the same claim on their FB or Twitter page/feed”

# [Pat Hayes] Vinny: curating sameness of claims done internally.

# [Tantek] Does “same claim” mean “same string”? Or is there some affording for variance beyond whitespace?

# [Pat Hayes] Vinny: sameness means can be substituted validly. ( agrees with Leibnitz)

# [Pat Hayes] Vinny: I pay lawyers and editors so we have to be careful about exposure.

# [Pat Hayes] Vinny: Need to establish mutually beneficial/trusting relationships.

# [Pat Hayes] Vinny: Legal liability means must be careful about transmitting information from private to public areas.

# [Pat Hayes] Vinny: three main points: disjoint editorial from data claim process is dangerous.

# [Pat Hayes] Fingerpointing = identifying source is OK in story, but maybe not in metadata. One doesnt know what will be done with it downstream.

# [Dan Brickley] When I met with FullFact various times, they often emphasize this point too, and argued that should highlight “expires” metadata for fact checks, that there is serious risk in having this sort of data floating around

# [Tantek] Why “expires”? scientific hypotheses don’t come with an “expires” date

# [Dan Brickley] @tantek the review of the claim as embodied in a document, … so ‘expires’ in the sense of the document carrying a “best before” date

# … so the expiration of the factcheck, not the fact

# [Pat Hayes] Vinny: third: ambiguous claims might not have clear sources or multiple (too many) sources Cant point fingers.

# [Pat Hayes] Vinny: satire is always false.

# [Dan Brickley] “Employment is falling”

# [Pat Hayes] Facebook don’t like satire.

# [Tantek] what are the stats on sharing Onion articles on FB?

# [Pat Hayes] Well, that certainly happens

# [Pat Hayes] Vinny: different attitudes to satire between us and FB can cause a shit storm.

# [Sandro Hawke] danbri: fullfact needs to be balanced, as a british charity

# [Pat Hayes] Vinny: Claim review data might be sensitive, time-dependent. Need contracts to ensure updates get transmitted to original source. Sharing is not simple, needs contracts.

# [Pat Hayes] DanB: Does this generalize?

# [Pat Hayes] Vinny: look at legal implications, historic firestorms due to misunderstandings and dataentry errors, then conclude its a general issue. Real costs involved, so revenue streams need to be considered also.

# [Dan Brickley] I’m hearing “comercial opportunity” and translating it to “institutional sustainability” in my head

# [Cheryl Langdon-Orr] LOL

# [Sandro Hawke] thanks pat!

# [Vagner Diniz] This the end to me See you tomorrow.

# [Cheryl Langdon-Orr] Host ended the session?

# [Dan Brickley] Thanks vagner!

# [Cheryl Langdon-Orr] See you then…

# [Dan Brickley] Sandro is rebooting the video meeting - sorry for the drop

# [Cheryl Langdon-Orr] *thx for the scribing work *hug**

# [Sandro Hawke] Sorry - it listed no one on the call

# [Sandro Hawke] will be bringing it back

# [Cheryl Langdon-Orr] I did wonder

# [Cheryl Langdon-Orr] no you dropped us ;-) ( I am so offended [not])

# [Cheryl Langdon-Orr] nP will join again at the top of the hour for the next session

# [Christopher Guess] Sally is thanking us all for being here, saying she recognizes some of us. She’ll talk and then do a workshop tomorrow.

# [Christopher Guess] She’s giving a general introduction to the “Trust Project” which she runs.

# [Scott Yates] Sally says that she’s in the great state of COLORADO!!!!! Woo hoo!!!

# [Cheryl Langdon-Orr] Joining in a moment

# [Christopher Guess] The trust project is a network of organizations around the world

# [Christopher Guess] I’ve been told to reference the slides, I’ll pick it up afterwards for the conversation

# [Dan Brickley] Sally: Uh-oh! [NO_CARRIER]

# [Cheryl Langdon-Orr] We were continuous in the Zoom room odd

# [Scott Yates] internet went down in the room for a minute

# [Christopher Guess] I’ve been asked to give just an overview of each slide, so here we go.

# [Christopher Guess] Slide: questions that people had about their news

# [Christopher Guess] include “The agenda?”, “The Journalist?” “Diverse Perspectives” “What’s the Source?” and “let me participate”

# [Christopher Guess] Slide: showing examples from La Republica, The Economist and MIC

# [Christopher Guess] This is how the organizations answer the previous questions.

# [Christopher Guess] Slide: an example of the label being overlaid on the NYTimes

# [Christopher Guess] Slide: examples of oraganizations agreeing to use the trust indicators: google, bing, facebook & twitter

# [Scott Yates] Sally gives props to the incomparable @Dan Brickley for making sure that Trust Indicators will flow into the platforms

# [Christopher Guess] Slide: UT-Austin did a study showing that the indicators seem to actually be helpful for readership

# [Christopher Guess] Slide: How we might communicate and work together with the trust project and other attendees’ work

# [Christopher Guess] Specifically on getting people to implement the work that the trust project has already accomplished and build onto it.

# [Scott Yates] Sally says that RSF and the JTI should work with bloggers because TTP is doing journalism organizations.

# [Christopher Guess] Slide: Thank you!

# [Christopher Guess] time for discussion

# [Christopher Guess] We’re going to talk about Trust Project first.

# [Dan Brickley] comes to mind… there may be a sufficient level of transparency where bad actors show themselves lacking

# [Christopher Guess] Sally: How do we asses them? That’s a good question.

# [Christopher Guess] Sally: “we’re learning the trust indicators are good at assessments themselves” They do a good job of showing what makes good journalism distinct

# [Christopher Guess] Sally: I worry about things that look like news providers on the internet but aren’t. Ex: police and local government sites, even universities

# [Christopher Guess] Sally: How do we do quality control afterwards? Now: manually

# [Sandro Hawke] Sally: We do vet everybody who wants to come in

# [Christopher Guess] Sally: Some of these organizations we’re collaborating with do their own assessment, which we can build on

# [Scott Yates] Sally: One of the orgs they are working with is NewsGuard.

# [Christopher Guess] Sally: there are a few other things we have to do to scale. Perhaps crowd source solutions?

# [Dan Brickley] This reminds me of IFCN code and their signatory verification process, … versus anyone being able to publish raw fact-check markup (in claimreview etc.).

# [Cheryl Langdon-Orr] Excellent briefing on the project … Thank you @Sally… appreciated…

# [Dan Brickley]

# [Dan Brickley] q+

# [Sandro Hawke]

# [Dan Brickley] can i squeeze in one more comment, to tie things back to Annette’s presentation earlier?

# [Scott Yates] (deleted)

# Christopher Guess says: white lists are small and the ‘net is big

# [Dan Brickley] w.r.t Chris’s point about scale, e.g. cf

# [Dan Brickley] is google’s view of daily mirror ‘about us’ page

# [Dan Brickley]

# [Dan Brickley] { “@context” : “”, “@type” : “NewsMediaOrganization”, “name” : “Mirror Online”, “ethicsPolicy” : “”, “masthead” : “”, “missionCoveragePrioritiesPolicy” : “”, “diversityPolicy” : “”, “correctionsPolicy” : “”, “verificationFactCheckingPolicy” : “”, “unnamedSourcesPolicy” : “”, “actionableFeedbackPolicy” : “”, “foundingDate” : “1903-11-02”, “ownershipFundingInfo” : “”, “diversityStaffingReport” : “” }

# [Sandro Hawke] Sally: TP isn’t approving the site, just that the site has agreed to disclosure

# [Dan Brickley] For Daily Mirror in UK, see also Wikipedia’s - their “inception” property does corroborate the “foundingDate” field asserted on the Mirror’s own site.

# [Pat Hayes] Is it possible to have a visual marker which is safe against backpacking? Eg that when clicked, performs a check and verifies to the user that is is OK (and if not, that it is not)?

# [Sandro Hawke] how to stop folks impersonating TP membership? maybe platforms will disallow that

# [Christopher Guess] Olaf is going to present now

# [Cheryl Langdon-Orr] Olaf’s audio is excellent for me…

# [Cheryl Langdon-Orr] I would like to think that collaboration is indeed a key… Agree @olaf

# [Christopher Guess] Olaf: Reporters without borders is made up of journalists but has the mandate of a humanitarian ngo

# [Christopher Guess] Olaf: many images are legal, but can be used by propagandists and as weapons against journalists

# [Sandro Hawke] +1 olaf: beware of these systems being used for censorship, and even against journalists

# [Vinny Green] Fake news is now a dirty word - Snopes uses the term junk news instead.

# [An Xiao Mina] *Re: “” – quote Fake news is now a dirty word - Snopes uses the term junk news instead. *

# [Christopher Guess] Olaf: anonymity saves lives of journalists in the global south

# [Cheryl Langdon-Orr] These sentiments are in keeping with much of the Report here

# [Sandro Hawke] olaf: transparency is a problem in the global south, where anonymity saves lives

# [Vinny Green] *Re: “” – quote @Vinny Green except your shirt *

# [Sandro Hawke] olaf: How do you deal with misuse of label? What about RT? Can it be games?

# [Christopher Guess] Olaf: regarding RT, “can it be gamed, whatever we invent”

# [Christopher Guess] Olaf: ethical codes and professional norms already exist in journalism around the world. The problem is compliance, do people really stick to them?

# [Christopher Guess] Olaf: How can we translate these norms to code

# [Sandro Hawke] olaf: Labels like “Fair Trade” don’t work well, we hear

# [Pat Hayes] NOt to mention “organic”

# [Christopher Guess] Olaf: When it comes to compliance we now use “standard padding”. ISO scheme specifically

# [Vinny Green] *Re: “” – quote *

# [Sandro Hawke] olaf: So many ISO standards already in the DNA of many companies – compliance on many things – so maybe they can pick this up. Infrastructure for conformity enforcement, including auditing, is already there. For ISO this is all out there.

# [Christopher Guess] Olaf: Enforcing, not inventing, standards is the most important part

# [Christopher Guess] Olaf: ISO standards are self-regulatory, not government based “The opposite of the law”

# [Pat Hayes] *Any phrase like “ news" is oxymoronically self-defeating. We should just call it "fiction", or of course "myth" or "false".*

# [Tantek] Note: “Fake news” as a term is not going away: (the “Talk:” page is quite fascinating)

# [Christopher Guess] Olaf: The most important thing is to preempt legislation

# [Sandro Hawke] olaf: This is fully self-regulatory, nothing to do with gov’t. The opposite of a law. The motivation is largely to preemt legislation, to do it before the govt gets involved. We may be past that point now.

# [Christopher Guess] *Re: “” – quote Any phrase like “ news" is oxymoronically self-defeating. We should just call it "fiction", or of course "myth" or "false". *

# [Dan Brickley] I’m curious which specific ISO standards, if it’s existing groups. e.g. (“Performance criteria for authentication solutions used to combat counterfeiting of material goods”)

# [Pat Hayes] “lies” conveys implication about the motives of the publisher. We can avoid that by speaking only of the falsity.

# [Pat Hayes] But lets take this offline :-)

# [Christopher Guess] Olaf: The core element of our project advertisers are important as well, they are trying to engage. It’s awkward when human rights advocates talk to advertisers

# [Scott Yates] We may run out of time, so I want to make a pitch for myself, which is that my new project, the Certified Content Coalition, is working to be the solution to the problem @Vinny Green brought up, that anyone can game the tags, and that @Christopher Guess said about white lists being small and the internet being big. If we can build a certification that lives inside the HTTPS cert. That’s a solution that can really scale, and be much easier for the platforms to read because it lives at the Transport Layer.

# [Dan Brickley] FWIW the Glasgow Media Group made a new for themselves around “Bad News” since 1974, … with fairly nuanced content analysis (quantitiative, qualitative).

# [Christopher Guess] Olaf: wish list: 1. bridging the gap between content and technology

# [Christopher Guess] Olaf: wish list 2. engage the platforms.

# [Sandro Hawke] olaf: wish list – bridiging the gap between content and technology which is happening already here; engage platforms – we can see they spend to support projects and send people to speak, but huge amount of expertise, doesnt seem to impact product

# [Sandro Hawke] olaf: #3 bridge widening rift between EU and US, which I find regrettable and sad

# [Vinny Green] *Re: “” – quote We may run out of time, so I want to make a pitch for myself, which is that my new project, the Certified Content Coalition, is working to be the solution to the problem @Vinny Green brought up, that anyone can game the tags, and that @Christopher Guess said about white lists being small and the internet being big. If we can build a certification that lives inside the HTTPS cert. That’s a solution that can really scale, and be much easier for the platforms to read because it lives at the Transport Layer. *

# [Christopher Guess] *Re: “” – quote And this from the earliest days of newspapers: *

# [Cheryl Langdon-Orr] room audio is low again

# [An Xiao Mina] *Re: “” – quote *

# [Vinny Green] *Re: “” – quote room audio is low again *

# [Sandro Hawke] pat: US values free speech to an astonishing degree, speaking as someone who has lived in both US and EU

# .. there are going to be these fundamental differences in attitude, we have to accept that

# [Sandro Hawke] olaf: press councils around the world hold media accountable

# [Vinny Green] *Re: “” – quote @Vinny Green We could potentially help journalism organizations with their corrections policies that will be consistent with, etc. *

# [Cheryl Langdon-Orr] Thanks great session !

# [Vinny Green]

# [Zoë Triska] * Aviv, Credibility Coalition: Can we agree on scientific and systematic ways to assess the credibility of information and then apply that at scale? Goal is not to provide a credibility score but to provide the data that can then be used in a credibility score.*

# [Jon Udell] Aviv Ovadya,

# [Jon Udell] Aviv: What are various frameworks for credibility-focused inspection of content?

# [Jon Udell] Two broad tools: Evidence chain, reputation network. Focus here on the former (but they are interdependent).

# [Jon Udell] The publisher control axis: Logical coherence is internally controlled, search results and Twitter propagation aren’t.

# [Jon Udell] CredCo focus: What we can inspect, on creator sites, that’s under their control.

# [Jon Udell] The CredCo study validated that certain indicators do correlate with credibility.

# [An Xiao Mina]

# An Xiao Mina says: , re: CredCo phase 2, it’s formalizing the indicators and methodology created in phase 1.

# [An Xiao Mina]

# [Dan Brickley] * aside regarding potential role of libraries, I asked a library friend (Karen Coyle) recently for stats:

“See,Academic-Libraries,Public-Libraries,Community-Libraries,School-Libraries,Other-Libraries/Weight/Totals-by-Country IFLA = International Federation of Library Associations That gives you number of full-time library employees in the world - not all librarians but it’ll vary by country. (1.4M!)”*

# [Jon Udell] CredCo research community developing an annotation guide for librarians, j-school students.

# [Jon Udell] Output: A collection of stories annotated for credibility. Annotations divide into content-based (intrinsic, look inside the article) vs context-based (extrinsic, triangulate on the article from other perspectives).

# [Dan Brickley]

# [Dan Brickley] see also (“Advertising should never disguise itself as the MC of the page. Pages with Ads that are designed to look like MC should be considered deceptive.” etc.)

# [Jon Udell] *“Advertising should never disguise itself as the MC of the page”

MC -> Main Content*

# [Tantek] I’d like to see lifecycle and economic cost estimates for gaming specific indicators

# [Tantek] lifecycle: indicator is conceptualized, implemented, first evidence of gaming found, gaming saturation (indicator no longer provides statistically useful results)

# [Jon Udell] *Indicators that seemed to matter, based on original study:

Representativeness of citations IFCN validation Clickbait assessment of title Exaggerated/emotional tone Aggressive ads*

# [Dan Brickley] I (at least personally) have huge concerns about exagerrated/emotional tone

# [Dan Brickley] q+

# [Sandro Hawke] +1 tantek: it would be a big win to reduce limbic hijack

# [Sandro Hawke] vinny: it would be good to downrank manipulative language

# [Pat Hayes] Vinny: people paying for associations with sources should be able to choose their rankings.

# [Jon Udell] If all other things are equal between two stories, and one is more measured in tone than the other, do we rank the former higher?

# [Pat Hayes] Yes.

# [Cheryl Langdon-Orr] hard to hear again…

# Christopher Guess says: “12 young people killed in houston” vs “12 6-year old childen killed, still in their art smocks” that’s going to get a different human reaction

# [Dan Brickley] There’s also the dying metaphors / language change side of things. Language change may just be happening faster…

# Scott Yates says: once you identify the ends of the spectrum, that’s a win, even if we don’t know the middle

# [Pat Hayes] But “smocks” and “six-year-old” is not charged *language, unlike say “slaughtered”//”killed” or “horrifying”, etc.. Facts can be horrifying while expressed in accurate unemotional language.*

# [Cheryl Langdon-Orr] Bye until tomorrow (my tonight 0200)

# [Sandro Hawke] ADJOURNED

# [Sandro Hawke] dinner @ 7pm at Limón Rotisserie @ 1001 S Van Ness.

Minutes Day 2

# [Dan Brickley] REJOURNED (if that’s a word)

# [Annette Greiner] Scribenick:Annette

# [Reto Gmür] present +

# [Annette Greiner] Scott: quick talk about Certified Content Coalition.

# [Michael Golebiewski] Can the speaker get closer to the mic.

# .. used to be a startup guy, always said you have to spend 80% of the time explaining what the problem is. This time, we know. There are lots of fake news problems. Dan: points out this is international Scott: This is the kind of content that is weaponized, not intended to be informational.

# [Michael Golebiewski] That is a lot better

# [Dan Brickley] how is audio now?

# [An Xiao Mina] for SF office folks -> :coffee: has arrived

# .. clickbait, stuff that’s trying to undermine democracy or the belief that anything can be true. Slide about Russian hacking of elections.

# [Annette Greiner] Zoe: I would argue that Infowars tries to disrupt democracy. Scott: Yes, that would be one. Dan: Is the Brexit stuff part of that? Scott: Yes .. mentions an RSF meeting where folks from the Russian embassy showed up. These are standards that are not coming out for a few years, and already there is interest there. We need a solution that can’t be gamed—public key infrastructure.

# …I am a journalist, former policy official, startup founder, researcher for CableLabs. They know how to develop PKI.

# [Tantek] (above) thread re: Infowars: Tantek: Infowars has ads, I’d ascribe greed as a motive

# .. Slide about what a solution should look like.

# .. Slide about categories of solutions .. the one that meets all criteria, is CCC, works like https

# [Annette Greiner] Dan: can you give examples of conclusions somebody might draw`?

# [Annette Greiner] Scott: this will operate in the background, good chance the consumer never sees it. It gives confidence the publisher is publishing good content and they are who they say they are.

# [Annette Greiner] Scott: this is one of the many things getting assessed at page load for content. Dan: what if someone from Reddit is interested? Scott: Reddit would just ingest the signal. Tantek: like certified bookmarks

# [Annette Greiner] Ben: what about bloggers? Scott: one new guy with one new blog is probably not going to get certified, but it’s a process for anybody who is serious about the long haul

# [Dan Brickley] Scott: talked to

# [Ben Werdmuller] Correction: I didn’t ask about bloggers - I asked about new publishers, startups, and ventures supporting new voices. It’s an inclusion issue: incumbent media is predominantly white, male, and from a very narrow demographic. We surely don’t want to lock underheard voices out or prevent new ventures from being started.

# [Annette Greiner] Ben: there is an inclusion issue here, and we don’t want to lock out any voices. Zoe: Facebook is concerned about this too. Ed: the challenge here is thinking about how it scales. We’ll talk about reputation later, but the idea of a proxy can be helpful. Think about a vetting mechanism. Scott: points out that a certification can be a marker of a new voice worth listening to. Ed: gaming could be an issue? Scott: it will be a little bit of a hassle to get. I envision 2 or 3 dozen places that can do the vetting. .. would be part of the https certificate, closed PKI. .. we are not a CA. They are going to be object ids that live inside the cert. Dan: could multiple locations do the same thing? Scott: the place where you want to have diversity is in the journalistic review. Slide about agreements (Scott will post this) Slide about how we cooperate. Scott: I talk about w3c all the time. This is how we fit into the project.

# [Annette Greiner] Scott: the Trust Project could be a certification authority. People could see that a site has Trust Project tags and is also certified. Jon: part of the security model is that browsers are configured to accept CAs from certain authorities only. Scott: Yes. Pat: users could check the certification Zoe: but users will never do that. Scott: also working with IEEE, CredCo

# [Reto Gmür] Audio become very bad

# [Annette Greiner] Scott: CredCo could help do a first-pass .. it would be a 501c6 nonprofit Dan: a US entity? Scott: Yes Slide about board members Scott: this only works if we’re all in it together. Dan: how is government represented? Scott: none; we don’t want to favor any one countrie’s government .. I hired a lawyer who deals with this stuff. He says this is like the self-driving car industry. We haven’t launched yet, need to get a board, enable standards committee, hire staff Dan: certifiers have incentives Scott: it’s not a deal until you decide, right. Sandro: there are 2 separate functions here, determining who is good enough and publishing a white list. So when you are helping OLaf’s group, you are doing something different.

# [Annette Greiner] Scott: other white lists wouldn’t live in a certificate. This is an indicator that happens even before the page loads. Dan: it seems a shame if the two services are shackled together. Scott: I’m not married to anything

# [Annette Greiner] Zoe: maybe the biggest danger is releasing a list of everyone who’s certified. Scott: the list could become a kill list for a dictator. Another reason to do it as a cert, not a white list. Zoe: we wouldn’t need to release that kind of info. Scott: carts don’t live in any one particular database. Tantek: you would have to crawl the whole web to reverse engineer it. It’s just one more attribute of a cert.

# [Dan Brickley] (I asked whether Verified Claims WG effort is related; @Tantek argues that it’s irrelevant, )

# [Pat Hayes] discussion: publishing white (or black)lists can be dangerous in some nations.

# [Scott Yates] From the slides: What Should a Solution Look Like? Operate Globally Require no workflow changes for content creators Consistent with Free Speech and Open Web philosophies Slow down maliciously generated content before it spreads Work on both content and advertising Have meaningful compliance Scalable to anything transported over HTTP

# [Dan Brickley] I guess replaces it?

# [Scott Yates] *Agreements You agree to uphold the ideals of your own editorial mission statement. You agree to correct errors on the same URL that the error appeared. You agree to use only advertising systems that are certified by the CCC✓. You agree to abide by the decision of your JCA, unless you appeal. If you appeal, you agree to note that on the disputed URL. If you lose the appeal to the CCC✓, you agree to make the change. If you lose the appeal and do not correct the error, your Certification will be revoked. If you lose three appeals in any two-year period, your Certification will be revoked.

Also, just for the record, here’s the background on part of what CableLabs does to show that this CCC system may seem complex, but it’s actually small potatoes compared to the cable world:

“The CableLabs PKI (DOCSIS®, PacketCable™, and CableCARD™) operated by Kyrio is trusted by cable operators, manufacturers, and consumers globally. It protects over $100 Billion in revenue annually in an interoperable ecosystem of many hundreds of millions of units, with hundreds of network operators, over 100 manufacturers, over 1,000 certified products in 55 product categories, it supports a retail market, and has been continuously operational for 18 years. Nearly 2.3 billion DOCSIS cable modems have shipped since 2000.”*

# [Annette Greiner] Sandro: there’s lots of other whitelist providers. Scott: the attribute in the cert would say that it was issued by the Trust Project Pat: Why is it even mentioned there? Ed: If Sally becomes a bad actor, you want to be able to pull any cert they ever issued. Tantek: you can revoke an authority Pat: how is that done? Tantek: the browser knows

# [Scott Yates] Thank you all very much!!!

# [Dan Brickley] *

3.4 Retail The retail domain encompasses all things where there is an exchange of value on an individual level. This includes brick-and-mortar store fronts, web-only venues, and even person-to-person sales.

R.1 Address verification Francis has found the perfect pair of shoes. When processing orders, Giant Shoe Company wants to be certain that his shipping address is accurate (inaccurate addresses are very expensive in terms of customer service). They offer a discount for customers who make verifiable addresses available as part of the checkout process. Francis offers his certificate and gets the perfect shoes for even less than he expected. R.2 Adult beverages June goes to her local beer and wine store to buy a bottle of wine. She submits her identity credential that lets the liquor store owner know that she is over 21 without having to reveal her actual date of birth, her address, or her state ID number. 3.5 Professional credentials In many aspects of life it is important to know that entities are who they say they are, and that they can do what they say. Professional accreditation is one way of learning about the abilities of an entity. Being able to verify these credentials is essential to their value.

… this feels same ballpark except expressed at the level of individuals rather than editorializing organizations.*

# [Annette Greiner] Sandro: it’s clear there is some utility to knowing whether a source of info is legit other than by just being on a whitelist. We don’t yet know what that looks like. There is research on reputation, but we don’t have a connection to that work.

# [Michael Golebiewski] You have a really bad echo in the room

# [Reto Gmür] can’t understand

# [Pat Hayes] If its any consolation, the sound is pretty bad even for us in the room :-/

# [Reto Gmür] when Scott was speaking it was no problem,

# [Pat Hayes] Discussion of reputation. Has to be subjective, no single truth. Fox news vs Maddow.

# [Reto Gmür] scribing woul be helpful

# [Reto Gmür] who was speaking

# [Pat Hayes] Tantek: idea of a single reputation (because truth is unique) leads to Gwynth Paltrow, results in harm.

# [Tantek] leads to trusting GP for health advice - which can result in harm

# [Reto Gmür] thanks

# [Pat Hayes] Tantek: humans are imperfect, can be experts in some topics, trolls in others.

# [Pat Hayes] General agreement…

# [Annette Greiner] Zoe: the closest thing FB has done is around trust. WE used that to elevate posts. For new organizations, it can be problematic. Vinny: do you have a complicated model or a simple model? Sandro: most of us have a mental picture of that. Tantek: that’s a horribly simple model. There’s also context. Sandro: there’s no absolute reputation. It has to be from a root. Eventually there are conflicts, and that’s where the reputation map comes into play. Tantek: context-free describing of reputation is the same anti pattern that results in GP being trusted. That’s a bad idea. .. It’s a way of reinforcing the appeal to authority. Dan: agrees with Tantek .. there’s an awkwardness when you talk about people vs their accounts. A person like my mum who would easily install malware by accident couldn’t work with that approach. Sandro: the platform can inform about when you’re leaving your safe zone. Zoe: America doesn’t know who to trust. Tantek: humans are imperfect. Experts vs trolls, there are some topics where the trolls are the experts. Sandro: I’m talking about a “good faith human” Tantek: even for single humans, you can’t ascribe that axis. Ed: this is infinitely complex.

# [Tantek] single axis of reputation is not only unworkable, but leads to reinforcing “appeal to authority” fallacies

# .. The person walking along the Hudson and trolling, some of the context isn’t relevant. Past behavior may be important and may not be significant. Sandro: GFH saves crowdsourcing effort. Chris: That’s brigading- you get a bunch of trolls together, things like boaty- McBoatface.

# [Sandro Hawke] chris: wikipedia is the counter-example to everything

# [Dan Brickley] /me bbialb

# [Annette Greiner] Dan: you may not know somethings a bot or a human trolling. .. you can get some of that by examining an account, but it may be replicated 100 times. Pat: you can’t do this stuff at political scale

# [Pat Hayes] sandro invented “fake noise”

# [Annette Greiner] Sandro: claim review seems amenable to crowd sourcing, but it can easily get diluted by “fake noise”

# .. one of 25000 people who report this are not going to get sued. Vinny: the platform might. it’s not the finger pointing that matters but the harm that results from it. If a publication gets downranked because of it their business might suffer.

# [Tantek] Aside: there’s a Wikipedia page for the phenomenon of vote brigading in particular: but nothing for the broader “rallying a mob online” meaning of brigading (e.g. GG attacks on individuals)

# [Annette Greiner] Sandro: say people are doing his in plain language on Twitter, that data could be picked up by FB and Google and used in their rankings. Who would get sued in that? Vinny: it’s different when the sole purpose of the application is the root of that. Wherever the standard is facilitated will be the one to bear the brunt of the reaction.

# .. if no one knows, then it prevents that potential brigading. .. the Snopes partnership with FB has elevated the amount of scrutiny.

# [Reto Gmür] q+

# [Tantek] perhaps we can discuss from positive examples? E.g. I’ll offer Swarm/Foursquare, and people’s reviews of restaurants, especially friends. OTOH I distrust Yelp reviews.

# [Reto Gmür] can’t hear @Dan Brickley

# [An Xiao Mina] *Re: “” – quote perhaps we can discuss from positive examples? E.g. I’ll offer Swarm/Foursquare, and people’s reviews of restaurants, especially friends. OTOH I distrust Yelp reviews. *

# [Tantek] An Xiao Mina, for that reason I specifically *unfollow people on Twitter who have patterns of rudeness or trolling (even if they are otherwise a subject matter expert) because I don’t want to ascribe any “general” trustworthiness to them.*

# [Reto Gmür] now you’re muted

# [Annette Greiner] Sandro: what can we do? Tantek: positive examples of reputation handling online. Swarm for restaurants shows which of your friends have been there and which like it. That’s one example. Sandro: collaborative filtering Tantek: Yelp reviews don’t work (carrot and onion) Sandro: you could have a class that is a GFH, maybe some vocabulary that lets people share what they want about reputation. Dan: reputation for organizations may be better than about individual humans. Reto: I think one very simple reputation-system agnostic approach would be to come up with a vocabulary where you consider a person to be trustworthy in a domain. “I trust Dan Brickley on this subject”. Would not necessarily need PKI. Sandro: we would have to research how could that not cause more harm than good.

# [Pat Hayes] For purchases, I read the most extreme positive and negative reviews, but as much to evaluate the reviewer as the product, until I find either some firm data or a lot of agreement between reviewers. Not sure if this applies more generally.

# [Dan Brickley] Sorry Reto. I said that I’m more optimistic about efforts to provide information about organizations, than about people. That perhaps this is my view as a European and tech platform employee, but GDPR/privacy/security/safety makes reputational claims about people much harder than re orgs. And that Tantek’s (regarding e.g. restaurants), CCC re editorializing orgs, and discusion about Wikidata / list of all ‘real’ newspapers, are all consistent with this.

# [Cheryl Langdon-Orr] It would be interesting to get a ‘social scientist’ aspect on the why and how we as individuals establish and maintain a “trust” or reputation with another individual or entity online

# [Cheryl Langdon-Orr] How do you propose to dal with us on remote for the next session?

# [Sandro Hawke] We just decided NOT to split

# [Annette Greiner] Next F2F is 23-25 Oct in France.

# [Annette Greiner] +1 to Cheryl

# [Cheryl Langdon-Orr] * the next F2F meeting clashes unfortunately with my ICANN 63 meeting commitments in Barcelona so my remote participation will be very sketchy .,,*

# [An Xiao Mina] *Re: “” – quote It would be interesting to get a ‘social scientist’ aspect on the why and how we as individuals establish and maintain a “trust” or reputation with another individual or entity online *

# [An Xiao Mina] *for our Inspection section:*

# [Vinny Green] I’m taking over as scribe!

# [Vinny Green] Check out the for documents from An

# [Vinny Green] Looking at the new study from Cred Co -

# [Vinny Green] Working with Hypothesis on content indicators

# [Vinny Green] Focus on breathe of citations, does the article include more research and sources

# [Vinny Green] Calibrating Confidence - To what extent does their confidence in their claims seem justified

# [Vinny Green] Proportion (Exaggeration - Minimization) Question: Does the author use exaggerations, minimizations, or generally seems to represent situations and events in a proportional manner? If so, highlight the relevant sections

# [Vinny Green] Emotional Valence Question: Is the language extremely negative, extremely positive, or somewhere in the middle? If not in the middle, please highlight

# [Vinny Green] We are looking here:

# [Cheryl Langdon-Orr] audio is very variable

# [Vinny Green] Aggressive sharing beyond just social share buttons

# [Sandro Hawke] better now @Cheryl Langdon-Orr ?

# [Vinny Green] “share this breaking story” - social shares in the prose as opposed to the buttons

# [Cheryl Langdon-Orr] right now YES @Sandro

# [Vinny Green] Need to difference between tone and rhetoric when assessing content.

# [Vinny Green] An wants to talk about potential indicators and what do we want to capture in our vocabulary

# [Dan Brickley] the connection between this and what Trust Project (and hence called “ReportageNewsArticle” vs “OpinionNewsArticle” is interesting

# [Vinny Green] Indicators have been assessed against science focused content, we want to make the indicator properties work for other types the content like sports and entertainment

# [Vinny Green] Show the results of the studies and a link to the data for the Content Credibility Indicators Vocabulary (CCIV)

# [Vinny Green] Link here:

# [Vinny Green] The platforms used in the studies are desktop based - Check by Meedan, Hypothesis, Public Editor

# [Vinny Green] *Content based assessments are anonymized.

When the brand of the publication is visible, the credibility goes down across the board.*

# [Vinny Green] During the study, rate it on a 5 point scale before and after the annotation practice

# [Vinny Green] Believe the product and process of annotation in this way can be google for media literacy

# [Dan Brickley] (curious - for the science article case, are there interesting differences between articles that link to university Press Release pages, vs those linking to scholary articles? or worse, just say “Scientists say…”)

# [Vinny Green] Look to raise the cost of gamification

# ..ooO(Is “Dead duck display in Bristol Ted Baker store causes outrage” emotive language? or just not very newsworthy)

# [Vinny Green] Looking at exposing more data and statistics to allow for independent researchers to see the possible emergence of gamification.

# [Vinny Green] Treat the survey question as the specification.

# [Tantek] folksonomy:

# [Annette Greiner] We need to make the vocabulary more agile than most.

# [Tantek] also “Limbic hijack”, apparently Wikipedia prefers

# [Tantek] adding a redirect shortly

# [Vinny Green] Facebook & Google will want a scalable solution. Human annotators build datasets, datasets trains models for scale

# [Vinny Green] Intervening with neutral questions before a user distributes content.

# [Vinny Green]

# [Sandro Hawke] “looks like you come here often, do you want to leave tip about it?”

# [Vinny Green] Field study, annotators can be asked to annotate articles that are being shared by the real world annotators.

# [Vinny Green] Certain demographics, regardless of the geo will not feel safe answering some of these questions

# [Vinny Green] Need a definition of emotion to capture misinformation, but not quality content but might have an emotional tone because of their passion and interest in a topic.

# [Vinny Green] Outrage language - some groups use language on a topic that is not outrageous to the group.

# [Sandro Hawke] “Weaponized Content”

# [Sandro Hawke] “Information Terrorism”

# [Sandro Hawke] Vinny: Commoditization and Weaponization of Information

# .. I’m comfortable with that term for State Actors

# .. artificial velocity

# [Sandro Hawke] tantek: it’s about causing harm

# [Pat Hayes] velocity?

# [Vinny Green] Facebook has a definition for Manufactured Sharing

# [Dan Brickley] Scooby Do discussion,

# [Pat Hayes] The meeting has dissolved into verbal playtime. Just in case anyone was wondering.

# [Vinny Green] Euphemistic language is a good indicator

# [Sandro Hawke] eg “Meddling in the election”

# [Sandro Hawke] eg “scared to use a term like ‘lie’”

# [Sandro Hawke] vinny: Eg “lie” is a conversation-stopper

# [Sandro Hawke] vinny: Does “False” stop people from reading the check

# [Sandro Hawke] tantek: conversation stopper == limbic hijack

# [Vinny Green] “Claimed with evidence”

# [Vinny Green] Optimize against limbic hijack could contribute to euphemistic language and dog whistle.

# [Sandro Hawke] vinny: It’s a huge challenge that FB often makes posts into images now, so the text is lost

# [Pat Hayes] Surely extracting text back from images is within state of the art?

# [Sandro Hawke] sandro: Maybe FB can or will include the associated text with the image

# [Pat Hayes] dissent/descent

# [Dan Brickley] *On British tabloid-eze, a peculiar variant on English - a miniature reader.

“Miracle tot cheats death as 12ft wall collapses” -> “If you are British and have grown up with tabloid newspapers, you will, of course, immediately understand that somewhere in Britain, a young child (a ‘tot’) recently managed not to die (‘cheated death’) when a high wall fell on or near them. You would understand from the phrase ‘miracle tot’ that the small child was extremely lucky indeed.” on “God mob burn TV licences”: No one would ever say “God mob”. “"”Has anyone overheard a person referring to someone else as a love rat? Can anyone recall saying that two people have set up a love nest? In the unlikely event that they did, would they then be said, in real life, to have indulged in a sex romp? Love rat. Love nest. Sex romp. These are common phrases in tabloid newspapers. Yet they form no part of normal daily conversation. … In tabloid-speak, girls are stunners, doctors are medics, police are cops, children are tots and, sometimes, kiddies. Sports stars are aces.

“ The orthography is recognisably English, but the order is all wrong; the tenses work differently, and some of the words – well, they’re in the dictionary, but that’s about the only place you’ll find them. This is because headlines don’t use English at all, but a language all their own.”

“Articles and possessive pronouns are practically nonexistent”; “All forms of the verb “to be” are similarly superfluous”; “The past tense is replaced by the more concise present simple”; “The future tense, meanwhile, is rendered by the infinitive”; “Noun modifiers are rampant”; “prepositions are often coopted to do the work of verbs”; …


# [Dan Brickley] *textual images OCR is increasingly solvable

For example, “Tesseract.js is a pure Javascript port of the popular Tesseract OCR engine.

This library supports over 60 languages, automatic text orientation and script detection, a simple interface for reading paragraph, word, and character bounding boxes.”*

# [Vinny Green] Lots of tools for Optical Character Recognition, need to get them into the workflow and products for annotators

# [Vinny Green] Read An’s book on memes and other things coming in January.

# [Scott Yates]

# [Dan Brickley] Facebook launched some great image accessibility work a little while back ->

# [Cheryl Langdon-Orr] Greg has had his hand up a while!

# [Vinny Green] Thanks Cheryl

# [Cheryl Langdon-Orr] I relly don’t require Intros as you all ashould know

# [Vinny Green] New doodle poll for scheduling the next meeting coming from @Sandro Hawke

# [Cheryl Langdon-Orr] other than initial ones of course…

# [Scott Yates] Who is our chart angel?

# [Cheryl Langdon-Orr] re new doodle … Oh good but do remember we are not all US based

# [Dan Brickley] FWIW I made a test post to facebook of an image (I didn’t supply any text), and it automatically added HTML alt tag text of “Image may contain: meme and text”. fb_meme_coffee_test_1.jpg

# [Cheryl Langdon-Orr] Let us know here when you reconvene please

# [Pat Hayes] In an hour.

# [Cheryl Langdon-Orr] Thx @Pat but I guess the accent is on the *when ;-)*

# [Cheryl Langdon-Orr] What is happening?

# [Cheryl Langdon-Orr] If you told us anything before then please not it was all just a buzz of …well word salad in the audio stream

# [Cheryl Langdon-Orr] Did you switch Zoom rooms??

# [An Xiao Mina] we are getting started now

# [Cheryl Langdon-Orr] Thank you for the update… we were shall we say concerned/confused for a while

# [Ben Werdmuller] I’m scribe for the browser session. Apologies in advance. Please let me know if I miss anyone’s name or get names wrong.

# [Cheryl Langdon-Orr] Echo!!!

# [Scott Yates] Did that fix it?

# [Cheryl Langdon-Orr] better now Thx

# [Ben Werdmuller] Sandro: a lot of W3C is about the browsers. Is there some connection between what we’re doing about the browsers. Tantek said he was going to come. [Tantek: I’m representing W3C.] So basically the question is, hypothetically, if a browser vendor was like, we want to do something right now, what could they do that might be helpful?

# [Tantek]

# [Greg] I wouldn’t mind just having a browser window extension where people could answer credibility questions on a scale similar to we saw, then be able to see it in the aggregate, or going back to reputation seeing how a particular group does this.

# [Greg] +1 @Pat Hayes for keeping in the art

# [Ben Werdmuller] Not just a little blue key, and not a full interstitial. Something between the two. If it’s a negative claim review.

# [Tantek] Re: full interstitial, presumably you mean phishing and malware sites:

# [Sandro Hawke] Vinny: beward flagging as false, since not-flagged==true

# [Pat Hayes] People will make that inference, whatever we say.

# [Greg] What about a source sidewiki…I miss sidewiki

# [Greg] I can enjoy them with my cookie modals

# [Greg] good point on mobile.

# Ben Werdmuller says: we see a lot of browser extensions, but we don’t invest in any bc of a lack of real-world use cases

# [Sandro Hawke] vinny: maybe in mobile SHARE action there’s a chance to intercept

# [Greg] You need to target learning tools as well, look at hypothesis, browser extensions aimed at students take up way higher than journalist, think starting at adult learners might be too late anyways

# [Greg] I also believe you do not understand how truth is shaped until you start shaping it online. any browser provider could break off a tiny bit of storage and coach people to make me an about me page that also builds in markers of credibility.

# [Greg] Wow @snopes is a 70% mobile company

# [Ben Werdmuller] [A short discussion about chocolate chip cookies]

# [Scott Yates] It looks like they won’t be coming anytime soon:

# [Greg] I think browsers would be very afraid for the same reason @snopes keeps claim/review editors private

# [Greg] I mean Google left “True Story of Martin Luther King” in the top ten results for a decade..Cloudfare had to fix it for them

# [Greg] well now that drm is baked in…..

# [Sandro Hawke] Browser could block sites claiming to be members of trust project which are not

# [Greg] now there is a browser institial that could be built right now, on any image upload say, did you give credit to the artist?

# [Greg] Want a better reading audience build browser tools to make them better writers

# [Tantek] s/fraud/fraught

# [Tantek]

# [Greg] I am applying for a large federal grant hope to build a social reader with this claim and evidence built in

# [Ben Werdmuller] This doesn’t require any tech; it’s just something a journal could do.

# [Greg] Build a web extension or app that allow folks to fill in Argumentative Vee Diagram

# [Greg] Satire is such a higher form of writing, I love to teach kids to make hoax websites so they understand how to spot credibility markers

# [Greg] I am applying for a job so I cite the person I think is going to chair the search wrote the paper….

# [Vinny Green] Here is an example of evil “satire”

# [Sandro Hawke] makaela

# [Ben Werdmuller] Makaela: We can look at templates for content. For example, we could check Wikipedia for templates that it has marked as being a negative template. We could then have a mechanism that makes a request to Wikipedia to request pages are made for harmful templates.

# [Tantek] e.g. if you go to you can discover the negative template on “Alternative medicine, pseudomedicine and medical conspiracy theories”

# [Ben Werdmuller] Makaela: it wouldn’t be censoring. eg if you have a cert problem, you can still go through.

# [Ben Werdmuller] Mikaela: I wouldn’t worry so much about the trolls as the businesses. If Goop gets wind that they’re going to get a redirect notice, there’s now a serious business reason for them to take down Wikipedia.

# [Ben Werdmuller] Mikaela: they’ve already started a new channel.

# [Ben Werdmuller] Mikaela: they were banned from doing any more live streams. They started a new one but as of today they’ve pulled the second live stream.

# [Ben Werdmuller] Mikaela: they have one strike at YouTube. If they have three strikes they’re completely banned. They’ve banned the livestreams but not the channel.

# [Tantek]

# [Tantek]

# [Greg] getting really bad echo, long delay

# [Dan Brickley] possibly similar experience:

# [Ben Werdmuller] [scribe off]

# [Christopher Guess]

# [Dan Brickley] is pertinent

# [Tantek] Re: SMS recovery vulnerability (and frequency of being hacked)

# [Cheryl Langdon-Orr] Echo is pretty bad at times … Sorry seems to be my mantra today

# [Cheryl Langdon-Orr] Can someone at least *type here what is going on please…*

# [Pat Hayes] Sorry, we had a longish discussion about personal and professional security against possible threats, and made a conscious decision not to record it.

# [Cheryl Langdon-Orr] That we understood

# [Pat Hayes] Upshot seems to be, anyone in the group who gets any kind of threat, alert other members of the group.

# [Cheryl Langdon-Orr] you have completed that now and are looking to ?? next session?? Wrap up?? fading and leaving for your weekend?????

# [Pat Hayes] 10minute break, then we will reconvene.

# [Cheryl Langdon-Orr] Thank you

# [Pat Hayes] Sorry about the chaos. Some folk are leaving to catch planes.

# [Cheryl Langdon-Orr] Again NP as long as we get a tiny sense of inclusion with occasional and timely updates Not helped by the variable audio at times either I guess…

# [Pat Hayes] The sound in this place is bad even for us. It is basically a large empty concrete box.

# [Pat Hayes] OK we are back around the table.

# [Cheryl Langdon-Orr] Thx

# [Pat Hayes] Sandro: we are close to having a list of issues.

# [Ed Bice] sandro - we are close to having a list of issues around CR (claim review)

# [Ed Bice] Dan - we have made a squishy data structure and put it in front of fact checkers

# [Cheryl Langdon-Orr] I will type echo, just because it is a tradition now… We will persevere…

# [Ed Bice] thx Cheryl

# [Dan Brickley] *

meta-issues with helping of post-conference braindump,*

# [Ed Bice] Sandro: Ed is the worst scribe in this group.

# [Ed Bice] verified

# [Vinny Green]

# [Ed Bice] Dan: philosophers have argued about what a claim is for thousands of years.

# [Pat Hayes] Claim identity is one central issue we have located.

# [Dan Brickley] *”"”Says Barack Obama began his presidency “with an apology tour.”

Several times, Obama “has apologized for what he deems to be American arrogance, dismissiveness, and derision; for dictating solutions, for acting unilaterally … .”

President Obama “went around the world and apologized for America.””””*

# [Tantek] “would” vs. “wouldn’t” ?

# [Ed Bice] Sandro: has an idea - if you change a word in a claim then the diff is another claim.

# [Sandro Hawke] ISSUE: Claim Identity

# [Ed Bice] Pat: no, this is contextual - for some claims (like quotes) yes.

# [Dan Brickley] q+ re promise keeping / flip flop management

# [Dan Brickley] related,

# [Ed Bice] Sandro: the claim ID is derived from the word order identity

# [Pat Hayes] It depends what you are, er, claiming ABOUT the claim. If you say you are quoting, get that textually right. If you are paraphrasing, can be looser with exact wording, etc..

# [Dan Brickley] Pat believe the morning star sucks

# [Greg] The essay scoring folks probably have a ton of experience with machine learning claims and evidence

# [Dan Brickley] vs

# [Sandro Hawke] Sandro: keep all 70 proposed fact-check claims, AND relate them to the editorial decisions, that boils it down into the checked claim

# [Greg] Going to just listen and lurk and play legos with my five year old

# [Ed Bice] vinnie: we fact-check or contextualize around the claim - there is a lot of human decision making that goes into this. At Snopes - the claim in the database it is what you see on the front-end.

# [Ed Bice] haha - Greg we are playing ideas legos here

# [Sandro Hawke] Vinny: the claim is a string. any miss-match makes it different

# [Sandro Hawke] dan; you also need some real-word context

# [Ed Bice] DanBri: do you re-phrase claims so that you are not amplifying the ill effect of a malicious claim

# [Sandro Hawke] Vanny: we’ll write a claim to minimize harm, sure

# [Sandro Hawke] Chris: PolitiFact will stick to the original text, I believe

# [Sandro Hawke] vinny: We’d put that in the body

# [Ed Bice] Chris: if there is a direct quote they will not alter it

# [Ed Bice] Chris: maybe use asterisks

# [Greg] questions as the manifestation of claims.hmmmm

# [Ed Bice] vinnie: snopes is a different type of fact-checking - not looking at politicians statements, checking the rumors that arise

# [Ed Bice] not fact checking at the point of speech we are fact checking at the point it becomes rumor.

# [Ed Bice] Chris: Politifact has added an explainer below their rating - eg, ‘cherry-picking’

# [Sandro Hawke] Chris: maybe add to the schema the dek/explainer

# [Pat Hayes] Vinny: claim needs to stand on its own. So a direct quote might not need context, but a remark about (say) unemployment might need context to be meaningful.

# [Ed Bice] Vinnie: the claim itself has to be coherent (checkable?) on its own.

# [Dan Brickley] q+ multi-language

# [Dan Brickley] *consider

excerpt, “”"”One missile,” the Fars post said, “had a message written on it that said in Hebrew, ‘Israel should be wiped off the Earth.’ “”””

English language discussion of Farsi and Hebrew political speech*

# [Ed Bice] Vinny: fact-checking is largely about contextualizing

# [Ed Bice] Chris: Full Fact automated fact checking tool works well sometimes and other times not so much - claims, like the framing of a photograph, are often fixed by the observer/photographer.

# [Cheryl Langdon-Orr] Echo

# [Ed Bice] Vinnie: at this point there is a human in the loop at all points with the FB fact-checking

# [Sandro Hawke] ClaimMatching

# [Pat Hayes] Echo better??

# [Cheryl Langdon-Orr] slightly

# [Ed Bice] Chris: working on claim-matching at Reporters Lab - offering the journalists a first pass at what sentences might be claims - augmenting the system.

# [Cheryl Langdon-Orr] at least we can distinguish an actual sentenced thread (unlike before) you should maybe see if the recording picks all this interference up or not ( I hope not)

# [Ed Bice] Vinnie: we cannot solve this with code. but can make much faster.

# [Ed Bice] awww - sorry @Cheryl Langdon-Orr

# [Pat Hayes] Code helps humans rather than replace them.

# [Pat Hayes] :-)

# [Pat Hayes] DanB: Need to discuss multilingual issues.

# [Pat Hayes] Eg bilingual countries.

# [Ed Bice] DanBri: how to encourage multi-lingual fact-checks?

# [Pat Hayes] Order of translate//praphase might be significant.

# [Pat Hayes] claim might involve translation from original speech, but also claim itself might get translated. Two issues.

# [Sandro Hawke] chris: isTranslationofClaim

# [Pat Hayes] DanB: both are relevant for us.

# [Ed Bice] Chris: think about a graphdb structure for this - ‘is translation of claim’

# [Sandro Hawke] ed: translations might change the meaning enough that the two commmunities end up disagreeing about its truth

# [Dan Brickley] q+ to ask John about hypothesis annotaitons of video / audio, deep pointing into video clips … URI syntax?

# [Sandro Hawke] vinny: maybe let the fact-checker convey their language, and then translations be explicit, like a toggle to the original version

# [Dan Brickley] cf (from discussion w/ tv team)

# [Ed Bice] Vinnie: in the schema if you can define the source language of the claim and make this visible to the end user viewing the claim in alt languages.

# [Greg] there are plenty of qualitative research platforms that point to both the region and timestamp in a media clip, no reason this couldn’t be a url

# [Dan Brickley]

# [Dan Brickley]

# [Dan Brickley]


# [Vagner Diniz] bye. the end.

# [Ed Bice] Tantek: Claimster

# [Ed Bice] Ed: Fire

# [Sandro Hawke] Issue-1: Claim Identity

# [Ed Bice] DanBri: let’s talk about pointing into videocontent - like second 350 seconds into the Obama speech - range or startpoint.

# [Ed Bice] * Jon: The standard player has to support selection which it currently doesn’t.*

# [Ed Bice] Tantek: #t=10-20

# [Tantek] See real world independent examples here:

# [Dan Brickley] Click with care - this is potentially distressing content - but has’s collection of TV broadcasts from 911

# [Dan Brickley] eg see

# [Dan Brickley] start=11&end=14

# [Dan Brickley]

# [Dan Brickley]

# [Greg] Mozilla Popcorn did this so well for fun

# [Pat Hayes] Sounds like its a standardization and acceptance issue rather than a technical one.

# [Dan Brickley] for another day, closed captions / subtitles, etc…

# [Greg] you could layer pop ups on a timeline and place them on the screen with a time stamp

# [Greg] a lot of schools use this:

# [Ed Bice] Makaela: using the first person editor tools on youtube for third parties to insert fact-checks

# [Ed Bice] Jon:

# [Dan Brickley]

# [Dan Brickley]

# [Dan Brickley] sudden flashbacks to

# [Ed Bice] Vinnie: inject a third party signal at a point in a video better than flagging the whole video.

# [Sandro Hawke] deep fakes is a reason we’ll need this, the provenance

# [Ed Bice] Sandro: Deep Fakes is a reason we need this more.

# [An Xiao Mina] arguably already necessary, with clever editing:

# [Ed Bice] Vinnie: I have always had this theory that poor eyesight is a primary enabler for fake news.

# [Sandro Hawke] tantek: do a study on that!

# [Pat Hayes] An Xiao’s example is very interesting, because the publisher has consistently claimed it was satire, but it seems clearly intended to mislead. Is this fake??

# [An Xiao Mina] boundary between satire and false news is very blurry now, on purpose:

wrap up

# [Ed Bice] DanBri: suggests that we simply kick off an arms war between two neural networks.

# [Sandro Hawke] tanket: next generation of weaponized content

# [Ed Bice] Tantek: what is the next generation of disinfo warfare?

# [Sandro Hawke] pat: fake video

# [Sandro Hawke] an: already out there, misleading editing

# [Sandro Hawke] pat: I mean deep fakes

# [Sandro Hawke] vinny: one mark zuckerberg one, where he’s going to sell facebook

# [An Xiao Mina] 100,000 views on this deepfake video:

# [Vinny Green] Horrible deep fake of Zuck:

# [Greg] what has the highest chance of getting into your notification pane, that is the next target

# [Dan Brickley] re claims and where I don’t think we don’t want to go — @Pat Hayes

# [Greg] I was talking to my son, my apologies

# [Sandro Hawke] Jon: illusion of diversity

# [Sandro Hawke] tantek: a form of sock puppet

# [Annette Greiner] Very well written junk news

# [Sandro Hawke] old, well-known websites

# [Sandro Hawke] tantek: What’s the next generation of info-weapon

# .. for deployment in this election

# [Ed Bice] Jon: coordinated msm divergence in russia as the method for establishing chaos

# [Sandro Hawke] makaela: trolls looking at younger generation, eg on Instagram

# [Sandro Hawke] tantek: you can forward stories

# [Sandro Hawke] makaela: people use the discovery page, so they’ll find stuff, for sure

# [Sandro Hawke] jon: sleeper bots

# [Sandro Hawke] vinny: a bunch of what appeared to fake police departments and/or news orgs

# [Sandro Hawke] tantek: Manchurian Bots

# [Cheryl Langdon-Orr] Thanks to everyone who is leaving now…

# [Greg] Goodbye all Have fun

# [Tantek] See ya Greg!

# [Cheryl Langdon-Orr] Bye

# [Tantek] Great meeting everyone! Thanks for lots of thought provoking discussion. A lot of what we discussed depends on some degree of rational thinking (checking/questioning sources, fact-checking), which itself is obviously systematically under attack (anti-intellectualism). I’ll leave you with this article about how the “Friends” TV Series in the 90s itself normalized a degree of anti-intellectualism:

# [Annette Greiner] Slides for the science session: