Governments Race to Rein In AI Deepfakes as Elections and Misinformation Fears Mount

Governments Race to Rein In AI Deepfakes as Elections and Misinformation Fears Mount A global regulatory crackdown is targeting the malicious use of artificial intelligence (AI) to create deceptive "deepfake" videos and images, driven by fears that the technology could undermine elections and spread large-scale misinformation. South Korea has enacted the world's first comprehensive AI safety law, placing direct legal responsibility on developers and service providers to prevent harmful AI-generated content like deepfakes and misinformation [55923]. The law requires companies to implement safeguards and reporting systems. This move follows a separate investigation by South Korean regulators into Elon Musk's AI chatbot, Grok, over allegations it can generate sexually exploitative deepfake images [58673]. The urgency for such measures is highlighted by incidents worldwide. In Nepal, a deepfake video showing three top political figures forming an alliance spread online ahead of national elections, despite the meeting never occurring. Experts warn such fakes threaten to mislead voters and distort democratic debate [57866]. Similarly, a consortium of global experts, including Nobel peace prize winner Maria Ressa, has warned that swarms of AI-powered bots could be deployed to disrupt the 2028 U.S. presidential election by imitating humans and infesting social media to reshape public opinion [56357]. Other nations are following suit with regulatory efforts. India has proposed new rules requiring technology companies to identify and remove deepfake content from their platforms, as part of a global push to control AI-generated misinformation [12872]. The technical and social challenges of enforcing such rules, however, remain significant. The Dutch selection of "hallucineren" (to hallucinate) as the 2023 Word of the Year underscores growing public awareness and concern. The term is now commonly used to describe AI systems that generate convincing but false or misleading information [27480]. South Korea Probes Elon Musk's AI Chatbot Over Deepfake Porn Fake Leaders, Real Fear: AI Deepfakes Target Nepal Election AI Bot Swarms Could Sabotage 2028 U.S. Election, Experts Warn India Proposes New Rules to Combat Deepfake Threat World's First AI Safety Law Enacted, Targets Deepfakes and Misinformation AI "Hallucinates" Its Way to Dutch Word of the Year