AI Misconduct.com

A daily rap sheet for the age of “superintelligence”.

Updated Nov 3, 2026
4,402 items in the archive

Tuesday, February 4, 2025

8 items

Plate cameras misread, Christchurch parents fined for daycare drop-offs

Who did it: HYBRID RNZ

Kindercare's chief executive estimates up to 40 parents believe they were wrongly fined for all-day parking at Christchurch's The Landing car park while dropping off or collecting children. Contractor Parking Enforcement Services, a Wilson Parking division, conceded some licence plate camera misreads produced incorrect breach notices, declined to say how many, installed an extra camera and apologised for any inconvenience.

Australia bans DeepSeek from federal government devices

Who did it: HUMAN ABC News (Australia)

Australia has banned DeepSeek from all federal government computers and mobile devices after national security and intelligence agencies found it posed "an unacceptable risk" to Australian government technology. Agencies must remove and block DeepSeek products immediately, though staff may still use it on personal devices. Home Affairs Minister Tony Burke said the government was "country-agnostic" in such decisions.

Judge says Musk's OpenAI suit will go to trial

Who did it: HUMAN The Register

OpenAI and Microsoft filed motions to dismiss Elon Musk's lawsuit, which alleges breach of contract, unfair business practices and fiduciary failures over OpenAI's for-profit wing, with OpenAI calling it "a grudge in search of a lawsuit". Judge Yvonne Gonzalez Rogers called Musk's irreparable-harm claims a "stretch" but, citing Microsoft's close ties to OpenAI, decided to let the case go to trial.

NBI raid on Makati love-scam hub finds real-time AI face-swap

Who did it: HUMAN GMA News Online

The National Bureau of Investigation raided an alleged love-scam hub in Makati City, arresting at least 167 Filipinos and two Chinese citizens. Agents found automated phones and an AI system that superimposed realistic female faces on scammers during video calls with targets in Europe and the Middle East. One suspect said trust was built before pitching cryptocurrency investments; charges are being considered.

Australian lawyer referred to regulator over ChatGPT's nonexistent cases

Who did it: HYBRID Legal Cheek

An unnamed Australian lawyer who admitted using ChatGPT to summarise cases for an immigration filing has been referred to the Office of the NSW Legal Services Commissioner after Justice Rania Skaros found it cited "nonexistent" cases. Citing time constraints and health issues, he said the summary "read well", so he used it without checking the details.

AI-voiced fake Google support calls target Gmail accounts

Who did it: HUMAN Infobae

Scammers are calling Gmail users from numbers that appear to be Google's, using AI-replicated voices to pose as support engineers and following up with emails from a google.com address, the report says, aiming to trick victims into handing over a genuine multi-factor code. Google said the account used was suspended and that it never calls about passwords.