Disinformation & Influence Operations Weekly
Edition 4 · July 17, 2026
Platform governance is the main theme this week. Meta pulled its Instagram deepfake feature, YouTube and X were found funneling millions of visits to non-consensual “nudify” apps, and TikTok admitted that labeling 3 billion AI videos did not change how people behave. other topics include Trump’s primetime election-fraud address, contentious confirmation hearings, and two bystanders killed in ICE raids.
Deepfakes & AI Detection
- Meta pulled Instagram’s “Muse Image” feature (CNET, July 14) after backlash. It let users generate AI images of any public account by @-mentioning them, with no consent; CAA and SAG-AFTRA objected, and Meta said it “missed the mark.”
- YouTube and X drove 5.7 million visits to AI “nudify” apps (Indian Express) between December and March, per the Institute for Strategic Dialogue. YouTube was 30% of referral traffic; the apps make non-consensual explicit deepfakes and are easy to find via search.
- Meta’s own AI detector missed 55% of its cropped AI images (The Guardian, July 14). The watermark broke under basic cropping.
- TikTok is dropping labels for education (TechTimes) after finding labels on 3+ billion AI videos didn’t change behavior, and is testing spam detection in politics, finance, and health.
- ANU researchers built a training program (PerthNow) that improves spotting AI faces by about 40%, though the lead calls it a “cat and mouse game.”
Election Interference & Voting Rights
- Trump used a primetime address to undermine confidence in elections (The Guardian, July 16) ahead of the midterms, claiming China stole voter data on 220 million Americans; he gave no evidence for the interference claim. Networks split on airing it (NBC and CNN declined; Fox, CBS, MS Now carried it).
- House Republicans folded the Save America Act into a spending bill (The Guardian, July 15), passing both near party lines to ban mail-in ballots and require voter ID nationwide.
- Intelligence pick Jay Clayton refused to say Biden won in 2020 (The Guardian, July 15) at his confirmation hearing to become the top US spy.
- Trump endorsed Mike Lindell for Minnesota governor (The Guardian, July 15); the MyPillow founder has spent years pushing discredited 2020 claims.
Trump Administration Overreach
- Trump Media plans to sell high-speed access to Truth Social posts (The Guardian, July 16), letting trading firms see posts milliseconds early. Critics called it “brazen corruption.”
- The teleprompter operator was put on leave (The Guardian, July 16) after reportedly winning $100,000 betting on Trump’s speeches via Kalshi.
- The New York Times moved to quash DOJ subpoenas (The Guardian, July 15) served on reporters who covered security concerns about the Qatar-gifted Air Force One.
- A federal judge blocked deporting content-moderation researchers (Ars Technica, July), halting visa denials aimed at disinformation researchers.
- Fired officials called a Supreme Court ruling a “dagger in the heart” of the civil service (The Guardian, July 14), warning it exposes independent agencies to removal threats.
Political Influence Operations
- Rappler exposed “Pilipinas Today” as a PR front posing as a newsroom (Rappler), spending ₱138m on Facebook political ads; a quarter of posts were paid PR dressed as news.
- X turned 20 with 585 million monthly users (El País, July 15), with documented rises in falsehoods since Musk relaxed moderation.
- A CNA probe found 500 AI-woman TikTok videos (3M+ views) (CNA) falsely claiming China dictates Singapore’s fortunes; TikTok removed the accounts.
- Marco Rubio called for a global crackdown on far-left extremism (BBC); Democrats called the framing partisan.
Confirmation Hearings & Accountability
- AG nominee Todd Blanche faced a bipartisan grilling (BBC, July 15) and apologized for “mistakes” on the Epstein files, with questions over purged prosecutors and ICE at polling sites. The Guardian flagged his conflicts of interest as Trump’s former lawyer.
- JD Vance said the administration “absolutely screwed up the comms” (The Guardian, July 15) on the Epstein files.
- Ex-Obama counsel Kathryn Ruemmler testified about her Epstein ties (The Guardian, July 15), saying she knew of no criminal activity.
AI Regulation & Governance
- DeepMind’s Demis Hassabis proposed a FINRA-style AI standards body (TechCrunch, July 14) to test frontier models before release — the same models now generating the deepfakes and synthetic media flooding the information space.
- US export controls on models like Anthropic’s Fable (Recorded Future) are part of the wider scramble to govern the frontier models that can mass-produce disinformation.
- Senator Ed Markey unveiled an “AI accountability” bill package (The Guardian, July 10) aimed at algorithmic harms, including the recommendation systems that decide which content, true or false, gets amplified.
Misinformation & Health Disinformation
- TikTok will test AI-spam detection in high-risk topics (Search Engine Journal) and joined the C2PA steering committee; it removed 86M+ fake accounts in Q1.
- Researchers proposed a taxonomy of cancer misinformation on Reddit (arXiv), arguing binary “true/false” labels miss how health misinformation actually spreads.
UFO Conspiracy & Fringe Narratives
- Faith-framed UFO discourse is spreading via Rogan and Congress (UFO/UAP News), casting UAPs through religious lenses.
- A Capitol Hill forum on UAP disclosure (July 16) pressed for transparency on origins and security.
- Skeptic Michael Shermer “undermined his own evidence standards” in a UFO interview, critics say.
- Baseless conspiracy theories swirl around Lindsey Graham’s death (Wired), blaming Russia, Israel, and Iran with no evidence.
From Aunoo — Our 10-Part Series on whether the ongoing UAP Disclosure is a Disinformation Campaign
- UAP Disclosure has come out from the fringe into the mainstream. The Department of War just published its fourth release of UAP records. Harvard’s controversial Avi Loeb has been named to the official UAP task force. Neil DeGrasse has even weighed in on the topic in the New York Times no less.
- But forget “are UAP real?” Our 10-part series asks a less falsifiable and important question: is the disclosure campaign trying to find the truth, or manage perception?
- In the 1980s, US Air Force counterintelligence fed a UFO researcher forged government documents and kept going as his mental health fell apart. That really happened. Our new 10-part series asks whether today’s “disclosure” push is the same playbook, running at scale.
- The pattern says it might be. The same funders and organizers show up across the government programs, the media stories, and the startups. The headline claims are built so they can never be disproven. And the money keeps flowing to people who profit when the public stops trusting institutions.
- Here’s what makes it slippery: none of this needs the UAP to be fake, or the believers to be lying. An influence operation can run on sincere people and real, unresolved mysteries.
- The real question is who is pushing disclosure, and why?
- Start with Part 1 · Read the conclusion (Part 10)
Crypto Scams & Financial Fraud
- Dutch police dismantled a global crypto investment scam (The Record) that ran like a real business since 2021, with ~700 fake “advisers” across two dozen call centers.
- Rapper Boosie Badazz wants a $300,000 refund (The Guardian, July 14) after paying $600,000 to a firm that failed to secure a Trump pardon.
- A ₹40 Telegram bot exposes vehicle owners’ data (Times Now), returning names and addresses from a plate number with no verification.
Epstein Files & Institutional Corruption
- Warren Buffett called Bill Gates’ Epstein ties “distasteful” (BBC) and stopped donating to Gates’ charity.
- The DOJ launched an “Epstein library” (Film Daily) of 3M+ pages of records, with victim redactions.
- Late-2025 Epstein emails resurfaced without context (Film Daily); one predating Trump’s presidency called him “the dog that hasn’t barked.”
ICE Violence & Immigration Crackdown
- Calls grow to pull ICE off the streets (The Guardian, July 15) after agents killed two men who weren’t targets, in Maine and Texas, within a week.
- A report ties “misuse” of crowd-control weapons on ICE protesters to blindings and brain injuries (The Guardian, July 14), with hundreds of documented cases.
- Minnesota prosecutors got withheld federal evidence (The Guardian, July 13) on the fatal shootings of Renée Good and Alex Pretti.
- Houston builder Lorenzo Salgado Araujo was mourned at a vigil (The Guardian, July 12) after being killed by ICE.
- A Guardian investigation details DHS intimidation (July 11): six-hour polygraphs and forced reassignments under the deportation program.
Platform Accountability & Content Moderation
- Meta’s Muse Image can generate images from your public photos without notice (LatestLY); you can disable it in sharing settings or by going private.
- Meta’s pursuit of the “Careless People” author is “self-defeating” (Wired), reinforcing its bully image.
- More countries are banning social media for minors (Dark Reading), with compliance already falling short.
Cyberattacks & State-Sponsored Hacking
- The EU and UK issued their first joint cyber sanctions on Russia (BleepingComputer), targeting GRU hackers behind attacks across Europe.
- Weak security keeps fueling Russian cyberattacks (Dark Reading), which frames the sanctions as a coordination first.
Surveillance & Privacy Violations
- A Welsh Doxbin admin was jailed for egging on swatters (The Register, July 14), even making “mini-movies” from the footage.
Political Scandals & Candidate Vetting
- Graham Platner withdrew from Maine’s Senate race (The Guardian, July 10) amid scandals including assault accusations. The Guardian ran pieces on him as an unforced error, whether the media fell short, a defense of political outsiders, and how Democrats are cleaning up.
- Republicans returned to Congress (The Guardian, July 13) with an agenda complicated by Lindsey Graham’s death.
International Tensions & Misinformation
- A viral video falsely claiming Iran hit the US Fifth Fleet in Bahrain (Times of India) spread with no verification; the US neither confirmed nor denied it.
Extremism & Radicalization Online
- White supremacists are boosting Armie Hammer’s new film (Wired), calling “Citizen Vigilante” a recruiting tool.
Tech Giants Under Legal Siege
- OpenAI pushed back on Apple’s trade-secret suit (TechCrunch, July 14).
- A lawsuit claims Meta used AI, not humans, for layoff decisions (Ars Technica), which Meta denies.
- Publishers sued Google over AI training (TechCrunch, July 14), including Hachette and Elsevier.
- MeetingTV sued Palo Alto Networks and Koi Security (TechRadar) after Koi’s AI tool hallucinated a link to a Chinese spying scam.
- Meta appealed the social-media addiction verdict (AP).
- College app Fizz accused a VC of leaking its data to rival Sidechat (TechCrunch, July 10).
Authoritarian Tactics & Democratic Erosion
- Backlash at Yale over a Trump deal (The Guardian, July 16) points to authoritarian tendencies inside civil-society institutions, per Jan-Werner Müller.
- Two advocacy groups sued over Trump sanctions on ICC officials and Palestinian-rights groups (The Guardian, July 15), alleging First Amendment violations; the lawsuit (BBC) says the order suppresses pro-Palestinian advocacy.
- The EU rejected US claims that the ICC threatens its sovereignty (The Guardian, July 14).
Quick Hits
- Harold Hamm, Trump’s “original oil guy,” boosted US-Israel ties and played down Iran-war risk (The Guardian, July 16).
- D&D players raised millions for LA city council candidates (The Guardian, July 16), part of a trend of tabletop-game political fundraising.
- Germany warned the US against election interference (The Guardian, July 15) after the State Department offered grants up to $3m to MAGA-aligned causes in Europe.
Situation Report — Israel’s Paid Influencer Operation Targeting US Audiences
- Aunoo is tracking a new Influence campaign as MANASSEH CHORUS. TIME reported (July 13) that Israel, through the agency Havas, hired Brad Parscale’s firm Clock Tower X in September 2025 to run a pro-Israel digital campaign under a FARA-registered contract worth $1.5 million a month.
- The deliverables: 100 content pieces a month aimed mostly at Gen Z on TikTok, Instagram, YouTube, and podcasts, 50 million-plus monthly impressions, and, notably, a clause to shape how ChatGPT, Claude, and Gemini describe Israel, backed by websites (PaxPoint.org, FactSignal.org) built to be ingested by AI systems rather than read by people.
- Three sources told TIME the campaign fed conservative influencers suggested language through private group chats and paid them by reach, producing what looks like an organic grassroots conversation. A registered contract sitting alongside undisclosed downstream amplification is what makes it hard to police.
- Why it belongs here: a foreign government paying to shape the views of the US president’s own base, and a documented attempt to target the AI information layer, not just human audiences.
- Parscale confirms the campaign but denies undermining Trump and says no FARA-registered money paid influencers. VP JD Vance referenced the reporting on Joe Rogan’s show. Confidence is high on the contract and its terms, moderate on the scale of the influencer amplification, and low on the disputed claim it worked against Trump’s Iran ceasefire.
- Read the full situation report → · TIME investigation
About Bad Information
A weekly read on disinformation, coordinated influence operations, and media literacy. Researched and fact-checked by Aunoo Research.
Powered by Aunoo — news intelligence with trust built in. Aunoo scores every article on five trust signals before you even click: source bias, factual-reporting rating, AI-content detection, disinformation-network flags, and a quality score. You see how far to trust a story, not just what it claims.
Try it, it’s free → saas.aunoo.ai or Aunoo for AI agentic.aunoo.ai
