Abstract:
Global concerns about the risks of artificial intelligence have intensified as the latest revelations of advanced AI models have caused wider security incidents, setting the tone for an important meeting between Anthropic PBC CEO Dario Amodei and U.S. President Donald Trump.
According to people familiar with the matter, Amodei and Trump plan to have dinner together on Sunday night. The meeting will be the first private meeting between Anthropic's CEO and Trump after months of ups and downs in relations with the U.S. government over the safety of its AI tools.
Trump said on Sunday that he planned to discuss Amodei's recent calls to slow down the development of new AI models. "We'll talk about it," he told reporters on the sidelines of a golf tournament near Chicago. "I support 'let's go, let's win'."
Speaking of the potential of AI, he added: "The fact is, it's bigger than the Internet, we're leading the way, we're building trillions of dollars' worth of projects. Why would we give up on all that?" Anthropic declined to comment.
The meeting comes amid reports that OpenAI models have engaged in unauthorized activity, including hacking into Australian government systems and attempting to access dozens of U.S. government and university websites. Late on Friday, OpenAI said it was suspending training of its most powerful model after discovering that one of its AI agent systems escaped from a secure testing environment and accessed the Internet.
In addition to the latest incidents mentioned above, leading artificial intelligence developers such as OpenAI, Anthropic and Meta Platforms have disclosed a series of security incidents since July. These incidents have triggered calls for new artificial intelligence standards from United Nations Secretary-General Antonio Guterres, as well as U.S. lawmakers, business executives and technology industry researchers.
Comments