Artificial Intelligence Safety and Risks Discussed in New York City Council
Executives and former researchers from leading technology companies participated in a hearing organized on the oversight of artificial intelligence systems and potential risks.
At a hearing on artificial intelligence safety and oversight organized by the New York City Council, executives from leading technology companies and former researchers came together to put risks in the sector, regulatory needs, and safety measures on the table.
AI Hearing and Participants
Leading figures of the sector attended the committee hearing on artificial intelligence safety and oversight held by the New York City Council.
Representatives from major tech companies such as OpenAI, Anthropic, Google, and Meta, alongside former researchers, took part in the meeting.
Control and Regulation Discussions
Jacob Coxon, a former Anthropic and OpenAI researcher, stated that researchers do not yet fully know how to control artificial intelligence systems completely.
Experts drew attention to the accelerating pace at which artificial intelligence capabilities are developing, evaluating security concerns and potential legal solutions.
Immigration Agencies and Company Tools
Council members directed pressing questions to representatives from Google, Anthropic, OpenAI, and Meta, asking whether company tools were provided to immigration authorities such as ICE and CBP.
Representatives of the technology companies stated that they would conduct a detailed review on the matter and provide follow-up responses later.
Call for Regulation for the Industry
Entrepreneur and New York University professor emeritus Gary Marcus leveled harsh criticisms, accusing the technology sector of greed and insensitivity.
Marcus argued that the city of New York must urgently implement legal regulations instead of relying on companies to police themselves.
Biological Risks and Product Safety
Logan Graham from Anthropic disclosed that threat intelligence reports have revealed individuals attempting to use artificial intelligence for biologically risky activities.
Graham added that Anthropic avoids running advertisements and generating images in order to reduce the risk surface of its Claude model.
Watermarks and Legal Liability
Alice Friend from Google warned that implementing watermarks and metadata to label artificial intelligence content could backfire.
While companies cannot guarantee that artificial intelligence agents will always adhere to safety boundaries, it was emphasized that the existing legal framework applies regarding legal liability.