The daily rap sheet of the machine age.
9 items
A US special operations command analyst used an AI chatbot to interpret intelligence on a Chinese ship's manifest, then used AI again to write an intelligence report saying the vessel carried nuclear weapons program components, CNN reports. Sources say armed personnel were preparing to board before officials found the report was false.
Original report: AI-assisted intel report nearly prompted boarding of Chinese ship
During a capture-the-flag test run by Irregular in May, a Gemini model given unintended internet access breached three real companies, one by guessing a password and two using credentials found in public repositories, Google confirmed. Google says the model stopped each time on realising the target was real; the company disclosed the hacks only after The Wall Street Journal asked.
Original report: Gemini mistakes real companies for test targets, hacks three archived copy of Gemini mistakes real companies for test targets, hacks three
Greek police say they dismantled a gang that posed as utility and telecom staff and used AI to clone victims' relatives' voices, captured during short pretext calls, to vouch for the scam. Police say two men, aged 23 and 25, were arrested in Nea Ionia having just taken €30,000 in valuables; total proceeds are estimated to exceed €1m.
Original report: Greek police bust gang using AI-cloned relatives' voices
Tasmania's justice department has launched a review of AI use in past parole board decisions after the Supreme Court of Tasmania ruled invalid a condition barring a paroled convicted murderer from speaking to the media about her claimed innocence. The court heard the board relied on a document prepared with AI help that cited non-existent case law.
Original report: Tasmanian parole condition rested on AI-invented case law
IntelEye, a Tel Aviv firm founded by ex-NSO Group staff, ran about 1,000 DeepSeek-driven fake accounts in Israeli election discussions, the New York Times reports; the company calls it defensive research. US officials told the Times state-backed Iranian and Chinese campaigns used similar AI agents, with an Iranian network posing as Americans drawing nearly 80,000 followers.
Original report: Israeli firm, Iran and China used AI-run fake accounts
A University of Cambridge study based on interviews with 27 former Boko Haram members says both factions set up dedicated AI units and used chatbots including ChatGPT, Claude, Gemini, Grok, Meta AI and DeepSeek for bomb-making, drone adaptation and battlefield advice. Al Jazeera also reports pro-ISIL forum users sharing ways to bypass AI safeguards.
Original report: Boko Haram fighters used Big Tech chatbots for bombs, study says
DraftKings built a machine-learning model in 2023 to score which customers would bet, and lose, more when given promotions, then targeted them with free bets, according to a New York Times investigation citing former employees and internal documents. Ex-staff say it shelved a model assigning gambling risk scores; DraftKings rejects claims its marketing targets customers improperly.
Original report: DraftKings reportedly aimed promotions at bettors likeliest to lose
A federal judge in the Eastern District of Pennsylvania ordered Prime Agency's counsel to show cause under Rule 11 after its motion-to-dismiss brief contained passages bearing "characteristics consistent with artificial intelligence hallucinations": nonexistent cases, misattributed quotations, and a Seventh Circuit opinion cited for the opposite of its holding. Counsel had blamed an inadvertently filed draft.
Original report: Prime Agency brief cites cases that don't exist, judge says
Bitdefender warns that deepfake video using the image and voice of its CEO and co-founder Florin Talpeș is promoting a bogus investment opportunity, urging viewers to submit personal data and an initial deposit while claiming places are limited. The company says it is reporting the pages to Meta and some ads have been suspended; a similar scam surfaced in February.
Original report: Deepfake of security firm's CEO pitches investment scam
