AI Misconduct.com

The daily rap sheet of the machine age.

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

Tuesday, July 22, 2025

6 items

Northern Marianas court dismisses fake-citation appeal, declares litigant vexatious

Who did it: HYBRID AI Hallucination Cases database

On July 22, 2025, the Supreme Court of the Commonwealth of the Northern Mariana Islands dismissed with prejudice pro se appellant Zaji O. Zajradhara's appeal against Northern Marianas College, finding he cited nonexistent, 'hallucinated' cases, misrepresented others, filed over forty documents and attacked opposing counsel. It declared him a vexatious litigant needing permission for future filings.

Deepfake Thai PBS anchor and Anne Jakrajutatip pitch fake investment

Who did it: HUMAN Thai PBS Verify

A fake Facebook page ran a deepfake ad, viewed 12,000 times, in which a Thai PBS World anchor appears to interview Miss Universe organisation CEO Anne Jakrajutatip about a 'government-certified' platform where 8,000 baht could return up to 800,000. Thai PBS Verify traced both clips to genuine footage; the anchor says she has been deepfaked at least 3–4 times in a year.

GLOBAL GROUP ransomware deploys AI chatbot to pressure victims

Who did it: HYBRID Infosecurity Magazine

Picus Security says GLOBAL GROUP, a ransomware-as-a-service brand launched in June 2025 and assessed as a rebrand of Mamona RIP and Black Lock, greets victims with an AI chatbot that automates negotiations and applies psychological pressure, aided by an urgency timer. Transcripts show demands reaching seven figures, such as BTC9.5 ($1m).

For 3,800 yuan, Chinese GEO vendor promises chatbot recommendations

Who did it: HUMAN 36氪:关注互联网创业

IT Times, republished by 36Kr, reports a 'generative engine optimisation' vendor promising '100%' AI recommendation for 3,800 yuan by publishing 50 article links per keyword so models such as DeepSeek cite clients. A tester's invented brand reached the top three in DeepSeek, Doubao and Wenxin Yiyan; DeepSeek also listed out-of-town restaurants as Shanghai picks.

Texas court warns pro se relator over non-existent case citations

Who did it: HYBRID AI Hallucination Cases database

In In re Cao, a Texas court noted that the pro se relator's brief cited non-existent cases and put her on notice that she may face sanctions for citing false caselaw in future, pointing to the Texas appellate rule allowing sanctions for groundless filings. The AI Hallucination Cases database lists AI use as implied.

Pennsylvania magistrate warns pro se plaintiff over fake case law

Who did it: HYBRID AI Hallucination Cases database

On July 22, 2025, recommending dismissal of Victor Kholod's pro se suit against Nationstar Mortgage and Wells Fargo, Magistrate Judge Susan E. Schwab found he cited a nonexistent 'Lefkowitz v. CitiBank' ruling, misstated another case, and listed eight cases the court largely could not find. The judge cautioned him about generative AI and warned of possible Rule 11 sanctions.