We’re Tracking a Potential Twitter Botnet Targeting Tech Researcher Timnit Gebru

A Note from the Collective: We’re taking a short holiday break, so this will be our last newsletter for 2020. Expect to hear from us in the first half of January. Stay safe and see you in 2021!

---

Botnet-Like Accounts Appear to be Targeting Timnit Gebru

In last week's edition of hatemail, we detailed the ongoing Google saga revolving around the ousting of expert AI ethicist Timnit Gebru (@timnitgebru) by company management after writing an unflattering research paper concerning the company and predictive language models.

Since then, one of our collective members has discovered what appears to be an astroturfing campaign targeting Timnit.

All the identified accounts used computer-generated profile pictures from thispersondoesnotexist.com; ironic, given that site uses machine learning to generate life-like pictures of humans. All the accounts also claimed to be current or former members of the Google Brain team — the team that Dr. Gebru worked on. Other Twitter users came across these accounts, speedily bringing about the takedown of the disinformation campaign in progress.

What is the purpose of this kind of activity? Previous campaigns may hold the answer. In 2018, more than a dozen Twitter accounts with strange coordinated messages spent their time defending Amazon during a time when the company was under attack for worker conditions in their warehouses. The suspicion was that the accounts were a coordinated attempt to better the public image of the company.

It is possible a similar tactic may have been used by Google, who is not only under pressure from their recent firings, but also under intense legal scrutiny. Alternatively, Miles Brundage (@Miles_Brundage) at OpenAI points out that the views and language of the accounts seem to reflect the views of Pedro Domingos (@pmddomingos), an AI researcher whose only praise seems to come from inauthentic Twitter accounts.

For reference and future study, we have collected a sample of the suspected accounts party to this disinformation campaign here and here and here and here.

Nation-State Warfare in Digital Space

• [New York Times] On Sunday, the Trump administration acknowledged that foreign actors (largely understood to be Russian state intelligence) had hacked into multiple government networks, compromising security at the Treasury and Commerce departments. It has since been discovered that DHS, the Department of State, and the National Institutes of Health were also compromised in the hack.
• [Reuters] [The Guardian] Back in October, news broke that a Russian operation disguised as a far-right news site was targeting voters in similar fashion to the meddling that took place during the 2016 election. Though Sunday’s news makes it clear that nations are not only interested in spreading disinformation, Facebook announced on Tuesday that it removed almost 500 companies and pages linked to France and Russia for spreading disinformation in the Central African Republic.
• [Lawfare] Disinformation campaigns are now becoming a standard tactic in information warfare. Shelby Grossman and Khadeja Ramali recently wrote about disinformation in the Middle East, and how governments are increasingly turning to private firms to strategize and execute disinformation attacks.

A Reckoning for Big Tech and Online Hegemonies

• [Revue/The Markup] Editor-in-Chief of The Markup, Julia Angwin (@JuliaAngwin), comments on the real cost of refusing surveillance capitalism. By refusing site analytics from common industry solutions such as Google Analytics, The Markup seems to have racked up quite a bill by baking privacy-conscious alternatives. While we think it’s the right thing to do, it’s still a hefty price to be able to refuse ‘feeding the Google.’
• [The Atlantic] Atlantic Executive Editor Adrienne LaFrance (@AdrienneLaF) pens this excellent summary for all the problems social media behemoth Facebook seems to espouse. From the harboring of disinformation to its monopolistic business strategies, LaFrance’s characterization of Facebook being a “Doomsday Machine” becomes clear when its scale is thoroughly examined.
• [Bloomberg] Unsatisfied by Facebook being sued by 48 states and the FTC, Texas filed an additional lawsuit against Google on Wednesday, alleging that the tech giant colluded with Facebook to game the advertising market and shutter competition.
• [Knuckleheads] Knuckleheads is an online community drumming up attention toward a single issue within the Google antitrust case: Web crawling. Google has an unfair market advantage to web crawling in that few other services are able to crawl the web like they can without getting blocked. Knuckleheads’ proposed solution would be to regulate web crawlers, and provide fair and equal access to the activity critical to a web search.

On Our Radar...

• [TechCrunch] [Vice] PornHub announced this week that it would be removing all unverified content, which amounts to millions of user-uploaded videos. Last week, featured Nick Kristof’s wave-making expose about Pornhub’s high tolerance for hosting sexual exploitation and nonconsent videos. Though its recent announcement may sound like the popular porn site is turning a new leaf, some suspect the changes are a public relations response to Kristof’s investigation since PornHub has done extremely little to curb sexual exploitation on its platform.
• [New York Times] Earlier this week New York City lawmakers called on the NYPD to halt sex trade stings after a ProPublica investigation found that associated arrests are often made with extremely little evidence and disproportionally targets non-white people.
• [CNN] Amateur codebreakers recently solved the Zodiac Killer's cipher, which had remained unsolved for more than 50 years. The codebreakers designed their own software around decoding the cipher with great success.

Hate speech website: ladssociety[.]com

Who hosts: Wix

Today’s site is ladssociety[.]com. It is the homepage for the Lads Society, a far-right white-nationalist group based in Australia and New Zealand. The site was recently cited in the New Zealand government’s inquiry on the Christchurch attack as an influence of the terrorist responsible.

We have observed that Lads Society’s site resolves to an IP address hosted by Wix, a global hosting and services provider, at 185.230.63[.]171.