Anthropic CEO: Government Should Have the Power to Block New Model Deployments

Odaily Odaily reports that Anthropic CEO Dario Amodei stated that if new artificial intelligence models pose specific risks, the government should have the power to block their deployment. Amodei published a lengthy article on Wednesday advocating for mandatory third-party testing of AI models to assess their potential risks across multiple domains. He wrote that if AI is deemed to pose "unacceptable risks," then "the government should have the authority to prevent or contain its deployment." This is one of Amodei's strongest statements to date in favor of strengthening AI regulation. Amodei wrote: "I believe that, at least in this current phase of exponential development, the most appropriate analogy is cars, airplanes, or pharmaceuticals. These technologies are crucial to the modern economy, but if designed or used improperly, they can also cause mass casualties." Previously, Anthropic had warned that its released AI model, Mythos, possessed the ability to discover and exploit critical software vulnerabilities, leading the company to limit access to only a select group of partners. This week, Anthropic released a new version that removed related cybersecurity attack capabilities. (Jinshi)