AI Hallucinations Spark Global Crackdown as India Enforces 3-Hour Takedown Rule

AI Hallucinations Spark Global Crackdown as India Enforces 3-Hour Takedown Rule Governments and health organizations are scrambling to contain the dangers of artificial intelligence "hallucinations," as the technology's tendency to generate false information triggers new regulations and emergency inquiries [27480][81632]. India has enacted one of the world's strictest rules, mandating that social media platforms remove any content identified as AI-generated within three hours of being flagged [78935]. The regulation, which took effect in February, specifically targets deepfakes—highly realistic fake videos, audio, or images—and other synthetic media that can spread misinformation [12872]. The move places a significant compliance burden on tech giants operating in one of the planet's largest internet markets [78935]. The urgency of such measures was underscored when a major mental health charity launched an emergency investigation. Mind, which operates in England and Wales, began a year-long commission after a newspaper investigation found Google's AI Overviews had provided "very dangerous" medical advice [81632]. The charity warned that millions of people with mental health conditions could be at risk from harmful AI-generated guidance [81632]. The public's growing awareness of the problem was reflected in language itself. In the Netherlands, the word "hallucineren" (to hallucinate) was named 2023's Word of the Year for its new meaning: describing an AI system that invents convincing but false information [27480]. Language experts say the selection captures widespread concern over receiving authoritative-sounding answers from machines that are simply making things up [27480]. Critics of rapid regulatory responses, like India's earlier proposed rules, argue that the technical challenges of identifying all deepfakes are immense and that the social complexities are not fully considered [12872]. The effectiveness of these global efforts to control AI-generated misinformation will depend heavily on their practical application and enforcement [12872][78935]. AI "Hallucinates" Its Way to Dutch Word of the Year India's 3-Hour AI Takedown Rule Begins India Proposes New Rules to Combat Deepfake Threat Mind Launches Emergency Inquiry After AI Gives "Very Dangerous" Mental Health Advice