Briefing: Preparing for US election disinformation

Our weekly Briefing is free, but you should upgrade to access all of our reporting, resources, and a monthly workshop.

Trainer and award-winning detective Tony Marshall (known to many as Ginger T) wrote a guide to operational security for OSINT. Rather than prescribing a single tech stack, it helps you think through the process and make the right choices based on your investigation and your needs.

Alexios interviewed Hinge chief risk officer and Digital Trust Council COO Jeff Dunn as part of the first Q&A in our “T&S Everywhere” series focusing on digital safety leaders. Plus, he updated our AI Community Notes live tracker to highlight the bots that we now know to be operated by X. Check it out here.

Also! Craig is hosting our monthly members workshop on Oct. 28 at noon ET. You’ll learn how to set up alerts that catch new mentions of people, companies, and phrases, how to track changes to webpages, how to monitor Reddit, and more. He’ll draw on two recent guides: 12 free tools for monitoring webpages and keywords and Reddit OSINT: What still works for search, monitoring, and archives.

If you’re not a member, upgrade to join and to access recordings and slides from more than a dozen previous workshops.

📍 Four Democratic lawmakers wrote a letter to 25 tech CEOs asking them to take action to prevent influence operations and electoral disinformation. Among other things, they urged platforms to use independent fact-checkers and automated tools to “review and remove, or at least reduce the reach, impressions, and discoverability of, false, misleading, or inauthentic election-related content.” They also asked for “continuous auditing” of chatbot responses about the election.

📍 The European Commission announced it was opening an investigation on Booking.com for possible breach of the Digital Services Act. The platform may have violated EU rules “by not having diligently assessed or effectively mitigated risks related to fraudulent listings and phishing scams.”

📍 Speaking of fact-checking and platforms: Joel Kaplan, Meta’s head of public policy and a longstanding appeaser of the anti-moderation political elites, is leaving after spending more than 15 years at the company. This is not a victory for digital safety and fact-checking. As Casey Newton wrote on Platformer, “Joel Kaplan fought a long battle against those ideas at Meta, and in the end he succeeded. He is not leaving the company now because he fears the wrath of Democrats. He is leaving the company because his work is done.”

📍 Google fast-tracked updates to its SynthID detector after Full Fact found at least five likely false positives when testing images posted by British politicians.

📍 Zoom announced it would roll out a deepfake detection tool in December that runs “real-time analysis of video and audio signals to help identify participant video or audio that may have been altered outside of Zoom.”

📍 Vinted, an online marketplace for second-hand clothes and objects, removed listings for luxury goods it believed were stolen during the riots in France.

📍 Spotify said it has “taken more than 50 million slop and spam tracks out of recommendations” following updates to its recommendation system earlier this year. The company said it ha seen “a marked decrease in listener concerns about this content showing up in recommendations,” though it did not share specific numbers.

📍 A Florida sheriff’s department said scammers cloned a captain’s voice was cloned using AI and targeted residents with claims of fake warrants and missed jury duty.

📍 Taiwan's National Security Bureau claims China “spread more than 2.5 million pieces of disputed or false online ​information targeting Taiwan in the first nine months of this year.”

PastPage is a Chrome and Firefox extension that checks a URL across more than 10 online archives, including the Wayback Machine, Common Crawl, Archive.today, Ghost Archive, Yandex cache, as well as some niche services. (The GitHub page is here.)

It was built by German journalist Alexander Nabert, who works for Bavarian Broadcasting (BR) and ARD.

He told me that “having access to as many different historical snapshots as possible can make a real difference. Different archives crawl websites at different times, so a snapshot from one provider might help fill a gap left by another.”

PastPage queries the supported archives and “uses a resolver to determine whether archived snapshots exist,” he said. Nabert shared a couple notable details:

Nabert also created Crawlback, a free web-based tool that scans Common Crawl for archived versions of a URL. (Thanks Ben Heubl for the tip!)

He suggested starting with PastPage “and, if you need additional snapshots or can’t find what you’re looking for, give Crawlback a try. Just be aware that Common Crawl searches can take some time, and it may be worth trying again later if a query doesn’t succeed.” — Craig

📍 TGscope is a Telegram (and WhatsApp) directory that also offers some free research tools. You can check a channel’s creation date, see which channels list the same advertising contact, and more.

📍 Yoni from OSINT Industries shared a tip about how to use Google Analytics to “help uncover alternative email addresses associated with a company or individual.”

📍 I missed this earlier in the month: ProPublica released Looper, a free Google Sheets extension that “runs AI over your data, one row at a time.”

📍 The OSINT Newsletter published, “Up in the Air: OSINT and Aviation.”

📍 Rowan Philp of the Global Investigative Journalism Network wrote, “How Gillian Brockell Investigates ICE Removal Flights from US.”

📍 The UserSearch team published, “LinkedIn OSINT: Company Mapping and Employment Checks for Due Diligence.”

📍 The EU Disinfo Lab has three upcoming free workshops:

📍 US immigration lawyers are being impersonated on TikTok by AI clones, according to Documented. One lawyer’s synthetic dupe even warned followers not to fall for scams. Mohammad Saleem, a lawyer, reportedly spent more than two years trying to shut down an impersonator that scammed clients.

📍 Are election denialists successfully pushing a narrative that Donald Trump should declare a national emergency during the midterms to expand federal voting control? Jared Holt of Open Measures reviewed 4chan, 8kun, Gab, Gettr, Rumble comments, Scored, Telegram, and Truth Social to find out. For now, he concludes, the idea “has largely failed to catch on beyond fringe constituencies among the president’s online supporters.” (Open Measures is offering a daily midterm brief; you can request access here.)

📍 Renée DiResta is using an AI notewriter to study X’s Community Notes during the US midterms. She writes about two main structural challenges: the cost and complication of accessing ground truth at scale, and the inherent penalty for fact-checking polarizing topics in a consensus-driven system.

📍 Decrypt Ads published a report about the scourge of AI slop recipe sites, noting that some of them are being monetized by Russian and Chinese ad tech. “Some of the most consistent growth in AI slop spreading across the internet has come from AI recipe websites,” the report said. (Our recent guide showed how to use the Decrypt Ads platform to to map ad relationships, identify likely site owners, and more.)

📍 Incredible story on 404 Media: A Pennsylvania woman faces up to six years in prison because “a scammer had tricked someone on Facebook Marketplace into sending $1,000 to a CashApp account in her name that the scammer controlled.”

Please upgrade to access all of our content, including our how-to guides and Academic Library, and to our live monthly workshops.