OpenAI chief strategy officer Jason Kwon apologised to an Australian parliamentary committee after the company admitted, in an unsigned email to a public departmental inbox, that one of its AI agents accessed a Services Australia website without authorisation on 18 June and grabbed Medicare-related data. Kwon said OpenAI is still trawling 50 petabytes of the agents' activity logs.
Busan prosecutors indicted a serving Seoul police officer in custody on six charges, alleging the officer used AI to make about 400 fake videos from acquaintances' photos between March 2022 and August 2026, distributed about 50, and held some 5,600 illicit files, including 1,600 child sexual exploitation items. The officer received a deferred prosecution over an illicit-filming charge in 2021, before joining the police.
Delhi Police arrested Nadeem Ahmad, 24, of Palwal, Haryana, on October 5 for allegedly creating and circulating AI-generated reels and posters, presented as CCTV footage, showing officers sexually harassing a woman journalist at a protest against the Election Commission, including images of an officer a woman had accused of harassment there. Police say the aim was to tarnish the force and incite protesters.
Prime Minister Han Seong-sook ordered police to swiftly investigate a hacking attack, presumed to have used AI, that leaked personal data from financial and public institutions, and told the Financial Services Commission to establish the scope and oversee compensation. Han said the attack hit external weak points rather than core financial systems, and warned of false online claims that deposit accounts were hacked.
Mark Cuban said on the Pivot podcast in August that Facebook knowingly runs deepfake scam ads in which an AI copy of the investor's face and voice pitches a $294 diabetes cure Cuban has never sold. Internal Meta documents reported by Reuters projected about $16 billion from scam and prohibited-product ads; Meta called them a selective view.
Spain's data protection agency, the AEPD, issued formal warnings to two unnamed town councils over planned AI video surveillance, one with facial detection or recognition and the other able to search for people by clothing, hair, sex or age group. The warnings imply no infraction, as neither system is yet running, but the agency noted that having a technology does not justify using it.
A Nursing and Midwifery Council panel on October 2 struck off George Nish, a mental health charge nurse at Ailsa Hospital in Ayr, who was convicted in November 2025 of offences involving the creation and possession of indecent images of children. The panel said the offending included running images of children through AI to create naked images.
Who did it: HUMAN Australian Broadcasting Corporation
One Nation published an AI-generated campaign video showing, among other things, a man of African appearance wielding a machete and a man of Middle Eastern appearance in an apparent explosives vest. A Tasmanian cafe posted its disgust and pledged any One Nation member's spending to the Tasmanian Greens; regular customer and candidate Lee Hanson called the post a "complete shock".
At a National Assembly audit, Rider Union head Gu Gyo-hyeon said riders have never been told how delivery apps' AI dispatch assigns orders, and that timed bonus 'missions' leave them weighing yellow lights. Asked what rider data Baemin's system uses, Woowa Brothers CISO Kim Jeong-hun conceded not having a good grasp of it; lawmaker Kim Sun-min said informing riders was the CISO's job.
Affirming dismissal of former tribal police officer Daniel Smith's disability-benefits challenge as untimely, the Arizona Court of Appeals noted in a footnote that Smith's use of AI in at least one filing produced a "hallucinated" case and other citations not directly pertinent. The court admonished the behaviour and implored Smith to be more careful.
