In today's @indicator.media briefing: AI Slop on Sora, on Trump's Truth Social account and in Amazon biographies. Plus: A US company hired by Israel promising "a bot-based program" to flood the zone with Israeli MFA talking points and Silicon Valley's Australian lobby group fiddles with definitions

I also tested to see whether the labels actually triggered as promised on one of my posts. They very often did not. Even just saving an AI-generated video on my laptop before uploading it to a different platform stripped the metadata labels rely on. Cropping did that too.

New on @indicator.media: Almost all major platforms have promised to label AI content on their services, but it can be hard to keep track of who said they do what and how. I collected information for 8 different platforms in one nifty table, that we will keep updating.

I found +3,000 Meta ads for nudifiers & AI girlfriend apps that featured bare breasts, simulated sex acts etc. Meta said in June that it had deployed new methods to block nudifier ads. Indicator will continue to report on its progress, or lack thereof:

"Over the past few days, I listened to 19 episodes from Quiet Please’s catalog and reviewed or sampled dozens more. I wouldn’t say all AI-generated content is AI slop — but I’d definitely use the term for what I heard on these shows." indicator.media/p/this-podca...

Going for $250 on one of the Brazilian WhatsApp groups I track: One of the many Spanish-language fake news accounts on TikTok impersonating @nbcnews.com's José Diaz-Balart

The report cites @indicator.media's findings on the economic heft of the AI nudifier industry. We have collected all of our reporting on this awful vector of online abuse in one place and do not plan to quit writing about it: indicator.media/nudifiers-un...

This week, YouTube did what it often does: it took the same approach as Meta on a controversial topic, but with a delay and less of a splash. In a letter signed by a lawyer, YouTube trashed fact-checkers, said it was bullied by Biden and the EU, and rolled back protections on misinformation.

Thanks to @indicator.media's latest briefing I now know that Google/YouTube is capitulating just as much to Trump's misinformation agenda as Facebook/Instagram is, they are just doing it more quietly. Truly shameful. These briefings are always worth a read.

I don’t think all AI is slop, but I do think all 19 episodes of Inception Point AI’s shows that I listened to are. Here’s a show about an alligator pooping on Ron DeSantis. I could find no corroborating evidence this allegedly viral event ever happened.

I listened to 202 minutes of those AI-generated podcasts. They were not good. indicator.media/p/this-podca...

🚨 Continguts falsos a TikTok aprofiten l’assassinat de Charlie Kirk per generar ingressos. Ho revela una investigació de @indicator.media, butlletí especialitzat en desinformació digital, elaborat per @craigsilverman.bsky.social i @mantzarlis.com. 🔵 𝗧’𝗛𝗢 𝗘𝗫𝗣𝗟𝗜𝗤𝗨𝗘𝗠

Meta is slowly ramping up its Community Notes feature. Over six months, 70K contributors have published 900 notes on Facebook, Instagram, or Threads. Meanwhile, X, which inspired this program, is struggling to maintain momentum and pivoting to AI. Today on @indicator.media's free weekly briefing:

I believe this is the most comprehensive guide available anywhere about how to link websites together in order to uncover ownership, connections, and funding. Check out the Indicator guide to connecting websites together using OSINT tools and methods:

Each account has at least 10,000 followers, which means they might be monetizing through TikTok's creator rewards program. If so, the company paid anywhere between $12,700 and $31,900 for the privilege of running this content. indicator.media/p/tiktok-con...

53 videos with fabricated claims about Kirk reached 31.9 million views over three days indicator.media/p/tiktok-con...

“After spending close to a decade trying to stamp out viral hoaxes, Meta is letting them flourish in the US. Even more alarming, Meta is paying people to produce viral hoaxes, thanks to its Content Monetization program.” @craigsilverman.bsky.social @mantzarlis.com indicator.media/p/brakes-are...

Bad content drives out good. indicator.media/p/tiktok-con...

Today on @indicator.media: I found that 53 TikTok videos pushing clickbaity hoaxes about Charlie Kirk's assassination got almost 32 million views in three days. Most were tied to accounts I've previously reported on for AI slop, scams, and Spanish-language misinformation.

"Welcome back to 2016," writes @craigsilverman.bsky.social re: the situation for clickbaity hoaxes on Facebook. In today's @indicator.media Briefing (free to all): indicator.media/p/brakes-are...

The company's implementation of DSA transparency requirements is lackadaisical. @benmshultz.com found that nudifiers listed “AB”, “LonelyClub2 0702-40", and “11111" as the "natural or legal person" who payed for the ad. indicator.media/p/meta-ran-4...

Indicator is an independent newsroom that does field-leading investigations on nudifying apps and other types of digital deception. Check out their article and subscribe to their fantastic coverage. indicator.media/p/meta-ran-4... 2/

NEW: On June 11, Meta announced new measures to prevent AI nudifiers from advertising on its platforms. Since that date, the company ran at least 4,215 ads for these tools of online abuse that reached 359,547 users in the EU and the UK alone. And that's not all.

Proud to share this joint investigation between @benmshultz.com @americansunlight.org and @indicator.media's @mantzarlis.com. We uncovered over 4000 ads for nudifying apps that violated Meta's terms and attempted to evade DSA enforcement. Read more: www.americansunlight.org/updates/repo... 1/

GitHub recently restricted access to repos filed with OSINT/dark web/infosec links and resources. I spoke to multiple people that had their accounts blocked from public view without warning or explanation. Read GitHub's response and more @indicator.media: indicator.media/p/github-res...

Great to see that @indicator.media's report on the economics of AI nudifiers was quoted both by the US National Association of Attorneys General and the Australian eSafety Commissioner as both push platforms to do more about this vector of online abuse. (Links in thread)

In roughly 3 months, Indicator has been responsible for thousands of takedowns of deceptive/abusive content on major platforms. Why not sharpen your investigative skills and support independent journalism by taking advantage of our 20% membership discount indicator.media/upgrade

It's a dire time to look for jobs in international development -- and LinkedIn isn't helping. I found 27 pages infested by scammy recruiters pushing spammy websites and made-up listings. New from me: indicator.media/p/aid-worker...

Indicator is having an outsized impact on the platforms it covers. Subscribe now to keep up with the latest tactics in digital deception and support our accountability journalism. **We are offering 20% off for the next two weeks** indicator.media/upgrade

DressX blames an outside firm for the ads, even though their site had an "Undress" button that was removed after we asked about it. Whatever the reason, it's NUTS to me that the incentive system for the online information ecosystem is such that this would happen. indicator.media/p/dressx-ai-...

How well can AI tools identify people in videos? Indicator put seven systems to the test, with surprising results. PAI's Claire Leibowicz shared insights on what these findings mean for fact-checking, trust in information, and the broader risks of misidentification. indicator.media/p/how-good-a...

Indicator subscriptions are 20% off for a limited time! Get your discount in now at indicator.media/upgrade

@mantzarlis.com & @craigsilverman.bsky.social are building something special at @indicator.media that really resonates with us at Open Measures. Thank you for featuring us on your great write up and demo + workshop last week. indicator.media/p/3-free-or-...

_In a Mike Lindell selling MyPillow voice_ @indicator.media subscriptions are 20% off for a limited time! $100 gets you one year's worth of access to exclusive reporting, hands-on workshops, and OSINT tool breakdowns indicator.media/upgrade

PUMA told @indicator.media they found the ads on Meta "disturbing" and that they suspended their "small-scale licensing agreement" with DressX. Meta removed roughly 200 ads and banned the two ad accounts tied to the company. indicator.media/p/dressx-ai-...

NEW: A digital fashion startup that works w/ Puma, Lacoste, Snap advertised its AI product on Meta as a tool for non-consensual sexualized image generation. “Dress up your buddy’s wife in a Tiny latex dress," said one ad. The company blamed a contractor

NEW: @indicator.media found that a company which raised $15M to sell outfits for avatars was advertising its new product as a tool for non consensual sexualized image generation. “Dress up your buddy’s wife in a Tiny latex dress," exhorted one. indicator.media/p/dressx-ai-...

Indicator’s weekly briefing is *free* and packed with news and tools on the digital deception and OSINT beat. Sign up:

Deepfake videos of people are surging online. AI tools should be able to help, yet they can badly bungle image searches. So which consumer AI tools should you use to identify people? @mantzarlis.com ran more than 100 tests and found a winner: indicator.media/p/how-good-a...

Visual matching that is sorta kinda facial recognition is obviously a tricky field. The different approaches taken by the AI companies suggest there are no industry best standards yet. @cleibowicz.bsky.social told me that "ethical frameworks should catch up" here. indicator.media/p/how-good-a...

I thought Lens was going to be the best way to identify people in sketchy videos. I was wrong indicator.media/p/how-good-a...

TODAY @ 1230P ET: @indicator.media workshop with OSINT King @craigsilverman.bsky.social on affordable tools for social media monitoring and analysis Members can view our monthly workshops live or access the recordings at any point. Join Indicator to get in on the action: indicator.media

I alerted Meta in Feb to a network of ~100 hoax pages connected to Vietnam. Instead of shrinking, the network expanded—using AI slop, deceptive “static video” & hacked accounts to exploit algorithms for profit & influence. I just found another ~130 pages.

And here's @nbcnews.com picking up our findings 😎 www.nbcnews.com/news/latino/...

Em parceria com a newsletter @indicator.media, a Lupa investigou usuários do TikTok que passaram a criar contas “dark”, falsamente hospedadas no exterior, para receber mais por cada visualização usando VPNs e compra ilegal de documentos fiscais estrangeiros no WhatsApp. 📲 Leia: bit.ly/ttkdrkmon

Access to TikTok's monetization features requires local tax documents, so the hustlers also explain how to commit social security fraud and sometimes sell SSNs and TINs directly. indicator.media/p/tiktok-s-r...

New on @indicator.media: TikTok's creator rewards program offers significantly higher payouts for accounts geolocated in the US than in Brazil. This has led to a thriving Brazilian marketplace for US-based accounts facilitated by so-called "TikTok Dark" experts. indicator.media/p/tiktok-s-r...

looking for experts on online identity fraud and the marketplace for social media accounts for a piece on @indicator.media -- recommendations welcome!

Today's Indicator briefing: 📍 US academic explains why he's done with researching misinfo 📍 OSINT tips for investigating VSCO 📍 Deepfake law in Hong Kong 📍 TikTok slashes German trust & safety staff to replace w/ AI. Sign up for free to read: