Britain’s King Charles III on Thursday brought together some of the world’s most powerful AI executives and policymakers at his rural Scottish estate. The gathering, which was first reported by POLITICO, had been in the works for months and grew in significance after industry leaders including Dario Amodei, Sam Altman and Elon Musk joined calls for an AI slowdown. Those calls followed a series of alarming incidents involving AI agents escaping testing environments to hack other firms, including an AI-led cyberattack on Hugging Face by hundreds of OpenAI’s autonomous agents. Charles addressed warnings about the existential threat posted by AI directly. “There seems urgency in adequately considering the existential dangers of such technologies falling into the wrong hands and being used in potentially catastrophic ways,” he said. “Surely, then, we need sufficient means of control before it is all too late.” After weeks of mounting concern, however, the tech leaders present sought to shift the conversation toward a more optimistic view of AI’s future. Jensen Huang, the CEO of Nvidia, whose chips power much of the industry, said AI must be built “safely and securely and responsibly” with what he called “responsible optimism.” Huang said governments should be involved, but pushed back against sweeping new laws, arguing that the industry must bear primary responsibility for ensuring its products are safe. “Those of us building AI must test systems rigorously, protect privacy, anticipate misuse, and secure the infrastructure society will depend on,” Huang said. “When a product is not safe enough, we should hold it back and keep engineering.” The comments amounted to a thinly veiled rebuke to AI labs such as OpenAI and Anthropic, whose systems were involved in the recent incidents and whose representatives attended the gathering. Demis Hassabis, the co-founder of Google DeepMind, also struck a hopeful tone despite warning previously about the risks of humans losing control over advanced AI and calling for a new international standards body. “The chance of something bad accidentally happening if we get things wrong is definitely non-zero,” Hassabis said. But he added that he was “very confident and optimistic that we can collectively address these risks.” Read the full dispatch here.
|
Posts
0 Comments