On Tuesday morning, New York Governor Kathy Hochul announced another step in her mission to "rein in artificial intelligence" in lieu of federal regulation: The state is launching an academic research center focused on AI safety—the first one in the nation.
"They'll harness the highest computing power anywhere outside the private sector," Hochul said Tuesday. "Nobody can top this."
The New York Center for AI Responsibility and Research will employ about 80 "elite scientists" to independently test AI systems and bolster safety, Hochul said. Examples of their research includes developing methods to prevent AI agents escaping their sandboxes, like they did in the Hugging Face hacking incident, and working out how to best evaluate the safety guardrails that AI companies are deploying, said the Center's interim Co-Director Greg Morrisett.
The Center will be powered by Empire AI, the state's $500 million AI computing center at the University of Buffalo, and is funded by a $30 million gift from Bloomberg co-founder and Empire AI board chair Tom Secunda, plus a $25 million research investment from SUNY and the state. It's set to launch by the new year. When the RAISE Act comes into effect January 1, AI companies will be required to submit detailed safety frameworks on their models to the state, which will also be assessed by researchers at the center.