Hastings urges government to mandate AI biosafety protections
Anthropic board member Reed Hastings argued that self-policing by AI labs is insufficient to prevent the creation of bioweapons. He called for mandatory government regulations to close the current regulatory vacuum before a catastrophic incident occurs.
Reed Hastings, a board member at Anthropic, has advocated for proactive government regulation of artificial intelligence, specifically regarding biosafety. Hastings stated that current self-policing measures, such as the classifiers used by OpenAI and Anthropic to prevent the design of harmful viruses, are inadequate. He argued that the government should require all AI systems to be protected against being used to design dangerous viruses. Hastings compared the current regulatory environment to the pre-2008 financial crisis, predicting that a significant incident will eventually trigger draconian regulations if proactive steps are not taken now.
He criticized the current approach to AI governance, noting that while AI leaders frequently meet at the White House, there is a lack of broader input from citizens and scientists. Hastings expressed skepticism about the political priority of AI safety, observing that the administration is effectively self-policing and that other matters, such as the construction of a ballroom, appear to be higher priorities. He also addressed public backlash against AI, attributing negative sentiment to fears of job loss and data center construction. Hastings suggested that demonstrating tangible benefits, such as curing diseases, is the only way to shift the public narrative. He emphasized that relying on voluntary industry standards is not enough to mitigate the risks of AI being weaponized for biological threats.