Heads of OpenAI and Google Back Call to Slow Down AI Development for Safety's Sake

Leaders of the world's biggest artificial intelligence companies have unexpectedly rallied around the idea of slowing the pace of AI development. OpenAI chief Sam Altman, Tesla and SpaceX founder Elon Musk, and Google DeepMind head Demis Hassabis have all publicly backed a call from Anthropic CEO Dario Amodei to ease off the throttle on AI development for the sake of humanity's safety.

It all began on Saturday, when Amodei published a lengthy message addressed to the entire AI industry. "We need to slow the rate at which we're improving the capabilities of frontier AI models. Progress will still feel breakneck, and that time needs to be used wisely," he wrote — referring to the need to close the gap between safety measures and the pace of technological progress, which he believes is widening. He wasn't calling for a halt to development, but rather for buying time to put proper safeguards in place.

Altman and Musk Unexpectedly Agree

Sam Altman wrote on X: "I agree with Dario that we need to slow the rate of capability improvement. This has been a major topic of discussion internally at OpenAI over the past few weeks." He went further than a simple show of support, endorsing a specific proposal from Amodei — bringing in independent evaluators with expanded access to companies' internal development systems. "Bringing in independent evaluators with employee-level access is a great idea. We'll be doing this too. More on this soon," the OpenAI chief said.

Elon Musk, true to form, kept his response brief: "Dario is right." With that, he signaled his belief that AI development needs to slow down, though he didn't specify what concrete steps his own company would take.

Rare Unity Among Rivals

On Sunday, Demis Hassabis of Google DeepMind joined the chorus of support for Amodei's call. He said the direction proposed by the Anthropic chief was the right one, though he stressed that the specific form safety measures should take still needs to be worked out.

Such unanimity is striking given the fierce rivalry among AI industry giants, who until now have raced to outpace one another at every turn. Up to this point, the tech companies' race to develop AI has resembled a superpower space race far more than a coordinated effort toward safety.

Source: novinky.cz