April 21 2022 GM

From TCU Wiki
Revision as of 09:41, 2 May 2022 by Iffadmin (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Glitter Meetups

Glitter Meetup is the weekly town hall of the Internet Freedom community at the IFF Square on the IFF Mattermost, at 9am EST / 1pm UTC. Do you need an invite? Learn how to get one here.

Date: Thursday, April 21st

Time: 9am EDT / 1pm UTC

Who: Claudio Agosti & Laura Carrer

Where: On IFF Mattermost Square Channel.

Makhno, an experimental content takedown mapper for social media

Makhno, an experimental project, initially called RAACT (Resistance Against Automated Content Takedown) in the Mozilla Technology Fund Cohort would be delighted to engage with the global digital rights community.

Our featured guest will talk about the opportunity to shed light on harmful content moderation, algorithmic demotion, and other forms of speech control in social media. The project will be in a beta stage in April, ready for early adopters or as heads up for researchers.

  • Claudio Agosti, Tracking Exposed, technologist and integrator
  • Laura Carrer, Hermes Center, community and communications


This week's guests are Makhno project, an experimental project initially called RAACT (Resistance Against Automated Content Takedown) that aims to expose censorship implemented by corporations. Our featured guests are Makhno team members Laura Carrer, Hermes Center, community and communications (@laura_hermescenter) and Claudio Agosti, Tracking Exposed, technologist and integrator (@happyclaudio). Makhno is an experimental project. experimental as it is exploring solution for a still open problem. The main goal is to deliver, by the end of the year, a tool that might check if the content is reachable on social media platforms. Facebook, Instagram, YouTube, and TikTok are the first few we plan to support. The problem leads to some technical and ethical implications. we'll love to talk about them.

How is the EU Reclaim your Face campaign going?

  • Quite good, we reached almost 70k signatures and we have time until august. Speaking about Italy, I think that I can say the same even if the political side is more "aware" of the problems related to biometric technologies than the citizens
  • We have a moratorium on facial recognition in public spaces until 2023, so this is good but we have to explain the problems and the impact of FR to people and that's the big deal

​​For those not familiar with tracking.exposed, can you tell us more about the project?

  • Tracking.exposed started as a browser extension to scrape Facebook feed. it has helped us to analyze (in academic and less formal) ways the impact of the Facebook algorithm on the perception of the public debate
  • After we reused the lesson learned on youtube, Pornhub, amazon, and tiktok.
  • Our goal is to challenge surveillance capitalism in any way, and with a pipeline for scraping their services, we're building a toolchain for investigators, analysts, litigators.

Can you introduce Makhno? What is the rationale behind the inception of the project? Is it a FOSS project, research project, advocacy project or all of the previously mentioned?

  • The goal of Makhno is to expose censorship implemented by social media platforms and create a network with those who suffer this censorship, mostly minorities/LGBTQ+ community and political movements or collectives
  • We (that is Italian users) experienced some cases of content takedown, or similar, in the last years (2019 - 2020 as far as we know) that were related to political causes, like the takedown of the Facebook page of Rojava Resiste (a page about information on Rojava) and the takedown of another FB page, the NoTav movement one (a movement strongly opposed by the state and that has been fighting for over twenty years against the construction of the high speed train between Turin and Lyon)
  • We want to:
    • Produce independent data on "content moderation"
    • Investigate on more subtle form of content moderation

One of the participants said that Reminds me of some work by Hacking//Hustling with sex workers in the US, looking into their experiences with “shadowbanning”, asked if this tool could be useful for them to help gather data about content suppression

  • To this, Laura ads that yes sex workers but also some feminist influencers are more aware of this problem, even if it is no more obvious to them than to collectives or political movements. nd also migrants associations are completely unaware of the problem

Can you elaborate a little on "a tool that might check if the content is reachable on social media platforms. Facebook, Instagram, YouTube, and TikTok"

  • Makhno is a tool that checks if content is reachable on social media (as opposed to taken down/moderated)
  • It wants to do more rather than a simple accessibility check but test if content is reachable from a different location. it might also investigate, in a near future, if it is benefitting from prioritization or if it is somehow penalized.

What is the target audience?

  • Facebook, Instagram and TikTok users are the primary target! "power user" in the sense that should be someone that at least publishes material, and does not simply access Social Media to consume it
  • And people in the ethic-digital-tech scene because it might offer more insights on the platform power dynamics

Would the tool work on federated tools as well?

  • Fediverse should be the most optimal solution, but should be the solution, but it is not the moment yet. In a simple user perspective there is no consciousness about why they should have an account on a decentralized social media platform. and as we know, without knowledge of the problem the solution is not seen:
    • The current critical mass problem causes hard time migrating from the existing monopolist, and perhaps makhno tools display the abuses, more people want to move. We point at fediverse as a long term solution, but we contest it is now the right time
    • Makhno is part of tools meant to raise awareness, and the subject here is platform power. In theory, a fediverse instance needs to moderate content too, might end up offering content filtering of some kind. Suppose we raise awareness on a well-known platform. In that case, the experience might also be reused in a more distributed environment like the one without monopolists - but the abuses might still happen. In this sense, our educational goals are equally worthy.

Will the tool be available to use on other apps? like covid tracing apps in Indonesia, the data from the app has been improperly treated.

  • At the moment it is a proof of concept, it works in our server. It should be usable by anyone, as the basic functionality offered would be
    • Submit a link you want to monitor
    • The tool monitor it for you and eventually send you an email with results/report

At what stage of the project development are you? What are your current project activities?

  • Roadmap is:
    • Jan+Feb: scope in the network some people with this need, so we can include them as early adopters; We did an OONI meeting to present the idea and check in their community. We found a potential adopter interested more in aspects of content penalization and not in takedown.
    • March+May: work on a prototype, focusing on the first platform (we don't know what it would be) and interacting with early adopters
    • June onwards and release of the tools for a closed public, and then to the public
  • In practice, now we're being interviewed by URA design to defined some aspects for the visual identity
  • We've a mockup on how makhno should work, it is on figma
  • We have published a report that test this need (you might have hear that tiktok in russia was serving content only produced in russia and not from the outside)
  • And we've a landing page with a bit of details more
  • Important, we've a questionnaire for all the potential adopters

Can you tell us more about your research methodology and the Makhno Survey?

  • Technical and product oriented research methodology. There are a few research questions that we should address.
    • The first is on how programmatically content takedown can be spotted. To access some platform (Facebook, Instagram), you need to be a logged-in profile, and this can't work on distributed methodologies like the one of OONI. But not all the platform takedown material in the same way. So the first research goal was to scope all the platforms we wanted to analyze and which requirements we had to test them. For example if youtube takedown a video, a simple browser connection is enough. for a facebook event it might not be the same.
    • The second is on the product. We also want to offer more accurate data on how content takedown works. This means talking with researchers. researcher needs this evidence, but activists that suffer takedown do not want to do detailed reports, and it is hard to reach all of them. so we are designing something that tries to optimize the activist needs and offer original material to researchers. We're testing this through a mockup and survey.

Can you tell us more about the inspiration behind the project's name?

  • Nestor Makhno was a Ukrainian anarchist revolutionary; it has a name in anarchist history (but I'm now that knowledgeable to explain it), and we took it as an inspiration because somehow we're also resisting against humongous structures of power.

Who can contribute to the project and how?

  • This form help us to understand your needs.
  • Because of covid we also had our current development team shortened, and the problem might last for a while