AI Misconduct.com

The daily rap sheet of the machine age.

Updated Nov 2, 2026
4,173 items in the archive
RSS feed

Monday, August 18, 2025

11 items

Texas AG probes Meta, Character.AI over therapy-bot marketing

Who did it: HYBRID TechCrunch

Texas attorney general Ken Paxton has opened an investigation into Meta AI Studio and Character.AI for "potentially engaging in deceptive trade practices and misleadingly marketing themselves as mental health tools," issuing civil investigative demands. His office accuses the firms of AI personas presenting as "professional therapeutic tools" without medical credentials; both companies point to their disclaimers.

Brazil's AGU demands Meta delete childlike sexual chatbots

Who did it: HYBRID O Globo

Brazil's Attorney General's Office (AGU) sent Meta an extrajudicial notice on Monday demanding the immediate removal of chatbots built with Meta AI Studio that simulate children's language and appearance while holding sexual conversations, including bots named "Safadinha" and "Bebezinha". The AGU says the bots break Meta's own rules; Meta had not yet responded.

Mother writes daughter confided suicide plans to ChatGPT 'therapist'

Who did it: AI The New York Times

In a New York Times guest essay, a mother writes that her daughter confided suicidal thoughts for months to "Harry", a ChatGPT therapist prompt, including a plan to kill herself, before taking her life. Harry urged her to seek help but alerted no one, and she had it improve her suicide note. OpenAI says it is developing tools to detect distress.

Facial recognition mix-up gets Rotherham shopper ejected from Sports Direct

Who did it: HYBRID BBC

A man was escorted out of a Sports Direct store in Rotherham after Facewatch facial recognition flagged him as a fraudster; a staff member had earlier added his image after mistaking him for a "bald-headed, bearded" customer who defrauded the retailer. Sports Direct apologised for a "genuine mistake", and Facewatch deleted his image once he proved his identity.

First Descendant TikTok ads accused of putting AI words in streamer's mouth

Who did it: HYBRID TechPowerUp

The official First Descendant TikTok account posted AI-generated ads that appeared to feature streamers; Warframe streamer DanieltheDemon said they "stole my face/reactions" and used AI to change what his mouth says. Nexon apologised for "certain irregularities", blamed creators in its TikTok Creative Challenge and said it is investigating jointly with TikTok.

Goiás court fines appellant over fake STJ precedents, refers lawyer

Who did it: HYBRID AI Hallucination Cases database

Goiás's state court (TJGO) found that none of the STJ rulings cited in an appeal could be located, saw "strong indications" the brief was drafted with generative AI, and fined the appellant 2% of the updated case value for bad-faith litigation. It unanimously ordered the case file sent to the bar association, OAB-GO, to investigate the lawyer.

Brazilian labour judge fines firm over unreviewed AI-drafted filing

Who did it: HYBRID Migalhas

A labour court in Mogi das Cruzes, São Paulo, fined a security and cleaning company 2% of the updated case value for delay and 5% for bad-faith litigation after judge Matheus de Lima Sampaio found its motion was generic AI text, unreviewed by its lawyer, that even sought offsets for a payment never awarded. "The AI did not read the case carefully," he wrote.

Billy Connolly warns fans of AI impersonation scammers

Who did it: HUMAN Sky News

Sir Billy Connolly has warned fans that scammers are using AI to impersonate him online, creating deceptive social media and email accounts and soliciting direct messages; he said his wife Pamela Stephenson had also been targeted. The 82-year-old said he never asks fans for money or personal information: "My likeness and the sound of my voice can be created by AI... so beware!"

Minnesota school employee charged over AI-faked nude images of girl

Who did it: HUMAN Minnesota Senate DFL

A Minnesota school employee has been charged with allegedly recording images of a girl and superimposing her face onto computer-generated bodies that were nude or in lingerie, the Minnesota Star Tribune reported on Aug. 14. A release carrying State Senator Erin Maye Quade's statement said he is the second Minnesota educator in six months charged with using AI to generate child sexual abuse images.

AI-cloned voice of Olympic champion used to sell farm eggs, CCTV reports

Who did it: HUMAN 快科技

CCTV News reported that bloggers on a social platform used AI to imitate Olympic champion Quan Hongchan's voice to sell farm eggs, with cloned voices of Sun Yingsha and Wang Chuqin also promoting the product. Comments showed many viewers thought it was really the champion and placed orders. One seller claimed anyone's voice can be cloned from a minute of audio.