South Korea launches 'AI Safety Watch', a standing citizen team to report generative AI errors and risks
The science ministry and the AI Safety Institute picked 80 of 138 applicants to report hallucinations, bias, safety-bypass and data-exposure flaws, and technical defects until August next year; verified cases will feed AI safety evaluations and be shared with companies, expanding a 30-day campaign that confirmed 464 of 625 reports.