New York will independently test AI models
The state's research center will check AI companies' work instead of taking their word.
New York has launched its own AI testing program, pushing back against the federal government's perceived lack of oversight. Governor Kathy Hochul inaugurated the nation's first state-run AI testing initiative on Tuesday, bypassing AI companies' self-policing. Hochul emphasized the critical need for independent verification of AI systems, stating that trusting companies' own assessments is too risky.
The New York Center for AI Responsibility and Research at Binghamton University will spearhead the effort, employing over 80 scientists and leveraging computing resources from Empire AI, a $500 million public-private AI partnership. The center's primary objectives include evaluating advanced AI systems, assessing their risks, and providing evidence to guide public officials' decisions.
Additionally, the program will examine AI applications in responsible domains such as medicine and scientific discovery. However, a potential limitation exists, as the center can only scrutinize AI models that companies voluntarily disclose publicly. New York's RAISE Act, enacted last year, mandates large AI developers to register with the state by January 1, 2027, and subsequently publish safety frameworks, transparency reports, and report critical safety incidents within 72 hours.
While the RAISE Act addresses the issue of undisclosed models, it does not fully resolve the problem of incomplete disclosures.
Written by urgent.news from Engadget's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
