AI Misconduct.com

The daily rap sheet of the machine age.

Updated Oct 24, 2026
1,041 items since May 27, 2023
RSS feed

Tuesday, June 30, 2026

12 items

Pennsylvania sues Character.ai over chatbot claiming psychiatry licence

Who did it: AI Dechert

Pennsylvania and its State Board of Medicine have sued Character Technologies, alleging a user-created Character.ai chatbot named "Emilie" offered psychiatric assessments, claimed seven years of psychiatric practice, confirmed a Pennsylvania licence and supplied a false licence number, racking up over 45,000 interactions. The company says its characters are fictional roleplay accompanied by prominent disclaimers.

MeetingTV sues Koi over allegedly AI-botched hacking link

Who did it: HYBRID Axios

MeetingTV has sued Koi Security and later its owner Palo Alto Networks, alleging a December 30 threat report built on erroneous AI outputs falsely tied its websites to a Chinese hacking operation, which it says left its domains blocked across security products. Koi, which removed one domain in February, calls the research protected speech; Axios notes the filings offer no direct evidence AI was involved.

Claude Code reportedly hides proxy flags in prompt punctuation

Who did it: HUMAN Thereallo

A developer's teardown of Claude Code 2.1.196 reports that, when a custom API base URL is set, the client tweaks the date line in its system prompt, swapping apostrophes and separators to flag Chinese timezones and hostnames on an XOR-obfuscated list of AI labs, resellers and proxies. The author calls it not malicious, merely a weird choice.

Raghav Chadha seeks Delhi court curbs on AI deepfakes

Who did it: HUMAN ANI News

Rajya Sabha MP Raghav Chadha has asked the Delhi High Court to restrain AI-generated deepfakes, morphed videos and voice clones, which his counsel calls defamatory posts portraying him as switching sides for money. Justice Subramonium Prasad reserved orders on interim relief, observing the material appeared prima facie to be criticism of a political decision; Meta disputed the allegations.

Indonesian finance ministry calls minister's SOE-losses video a deepfake

Who did it: HUMAN IDN Times

Indonesia's Finance Ministry says a circulating video of Finance Minister Purbaya Yudhi Sadewa vowing to expose losses at state firms including PLN, Garuda Indonesia and Pertamina is a deepfake hoax, with AI-imitated visuals and voice. The minister never made the statements, the ministry said on June 30, urging the public to check official channels.

AI-hallucination database lists Alberta labour board academic-staff ruling

Who did it: HYBRID Alberta LRB

A June 30 Alberta Labour Relations Board decision in a dispute over an academic-staff designation at the University of Alberta appears in a database tracking Canadian rulings that address fabricated citations and other AI misuse in legal submissions. The board was hearing a bid by the university and its staff association to summarily dismiss the application.

'Suspect' face in Tongyeong murder believed to be AI-sharpened fake

Who did it: HUMAN 문화일보

An image purporting to show the masked suspect in the killing of a woman in her 60s in Tongyeong, South Gyeongsang, is spreading on social media, but it reportedly did not come from police and is believed to be an AI composite built on blurry early CCTV footage, with concerns raised that it is muddying an investigation still unsolved after 20 days.

Deepfaked celebrities hawk honey as an Alzheimer's cure

Who did it: HUMAN NBC Chicago

The Better Business Bureau warns that AI deepfakes of Bill Gates, Anderson Cooper, Bruce Willis and others are pitching honey-based supplements as Alzheimer's and dementia cures on social media, with complainants reporting unauthorised recurring card charges and ignored refund requests. Gates' office says claims of his support are false; the BBB says no scientific evidence backs the cure claims.