Anthropic CEO calls for pacing AI development, setting guardrails

Emeritus professor Geoffrey Hinton, widely known as the "godfather of AI," has backed calls for artificial intelligence companies to slow down development, citing significant risks the technology could pose. Anthropic CEO Dario Amodei warned that rogue agents could take over the entire internet within six to 12 months. Hinton, a British-Canadian cognitive psychologist and computer scientist, agreed with Amodei's warning, stating that nobody knows how to estimate the probabilities of these risks.

Superintelligent AI is expected to arrive within a decade, according to most experts. Hinton argued that governments are moving too slowly in regulating AI and suggested implementing pre-release testing for chatbot models and obliging DNA synthesis firms to screen what they are asked to make. He acknowledged the potential for AI to cause irreparable harm through engineered viruses, mass manipulation, and damage to critical systems.

Despite advising a slowdown, Hinton emphasized that AI is currently doing tremendous good in various fields, such as radiology and drug design.