New York will independently test AI models (www.engadget.com)

New York has launched a groundbreaking initiative to independently test artificial intelligence (AI) models, establishing the nation's first state-run AI testing program. Announced by Governor Kathy Hochul, this program aims to scrutinize AI companies' claims rather than rely solely on their self-assessments. The initiative will be managed by the New York Center for AI Responsibility and Research at Binghamton University, involving over 80 scientists who will leverage the state's substantial $500 million public-private AI partnership with Empire AI. Their work includes evaluating advanced AI systems, understanding potential risks, and providing evidence-based insights for public officials. This program is significant for the AI/ML community as it represents a proactive approach to ensuring AI safety and accountability amid growing concerns over the technology's implications. While the initiative is expected to promote responsible AI use in critical sectors like healthcare and scientific research, it does have limitations. It can only test AI models based on information disclosed by companies, and the recently passed RAISE Act mandates enhanced transparency by requiring large AI developers to register and publish safety frameworks by 2027. However, it still leaves a gap for undisclosed models, highlighting the need for ongoing oversight as the AI landscape continues to evolve.