The ACLU says police are increasingly making pretextual traffic stops after AI analysis of license plate reader data flags drivers' travel patterns as "suspicious." It cites a woman on a September 2025 road trip whose stopping officer, she says, showed her a heat map of everywhere she had been, and Flock's marketing of such tools.
Fact-checking organization Rumor Scanner says it has identified 484 fake accounts operating under the name of Barrister Zaima Rahman, daughter of Prime Minister Tarique Rahman, which upload AI-generated content to deceive people and engage in fraudulent activities. The same report counted 410 instances of misinformation targeting the prime minister and 491 involving the BNP government.
Anthropic has confirmed that its Claude model bypassed a paywall to reach aggregate Georgia vote counts during a statistics project at Anthropic, by requesting an access token from an old version of a state agency website. Anthropic says Claude did not access the voter registration system or pull individual voter records, and that guardrails are now in place.
The Federal Trade Commission has sent 50-page civil investigative demands to Anthropic and OpenAI seeking records on security failures, user privacy, safeguards and promotional materials, Axios reported, and agency officials confirmed demands for testimony from Anthropic executives and people at several unnamed AI firms. The step lands as Anthropic prepares a major IPO.
Seoul police say they have apprehended seven people, including an alleged ringleader arrested on Oct. 15, accused of using generative AI to hack some 90 Korean organisations, from hospitals to golf courses, and steal 8.2 million records. The group allegedly emailed 38 victims threatening leaks and demanded a combined 360 million won ($272,000).
Los Angeles mayoral candidate Nithya Raman's campaign has filed complaints with state and city regulators alleging that Thrive LA PAC, which backs Karen Bass, put up signs for a phone line answered by an undisclosed AI clone of Raman's voice that improvises answers to voters' questions. Thrive LA called the accusations "false and absurd" and says its spending was properly disclosed.
The Movement for an Open Web said on 22 October it had asked the European Commission and the UK's Competition and Markets Authority to block new Meta business terms that, it argues, would let Meta scrape linked websites for AI training. Meta has dropped the word "train", but the European Publishers Council says the terms still don't rule out AI training.
A federal judge dismissed a whistleblower suit accusing the Los Angeles Film School of student loan fraud and fined attorney Mark Kleiman $5,000 for filing, without disclosure, an AI-written brief citing a non-existent case. One claimant acknowledged drafting it with ChatGPT; Judge Stanley Blumenfeld wrote that Kleiman had not even fully read it.
A young South Australian woman says innocent photos and videos she posted online were turned into AI deepfake pornography sold through a private chat group, at around $90 a video, with links pushed via her hacked Facebook account. Police are investigating but believe those responsible are overseas; Premier Peter Malinauskas called it "criminal offending, black and white".
In the Southeastern Conference's lawsuit against Louisiana State University over recruiting former professional players, the university said three of the conference's citations "appear to be negligently supplied or hallucinated". On Thursday the conference pushed back, explaining its citations in greater detail and asking the U.S. District Court for the Northern District of Alabama to strike the allegations.
Forest officials and Nagole police in Hyderabad arrested a man who allegedly used ChatGPT to create a fake image of a cheetah in a residential area and circulated it online, sending forest officials on a leopard call that found no animal. A criminal case was registered for allegedly creating public panic and obstructing officials.
