AI could be as dangerous as nuclear war, ChatGPT creator warns

Sam Altman, CEO of OpenAI, the creator of ChatGPT, expressed concern about the threat of artificial intelligence. Photo: ELIZABETH FRANTZ/REUTERS

The leaders of the world's largest artificial intelligence laboratories warn that artificial intelligence can be as dangerous as nuclear war.

Leaders and scientists, including Sam Altman, CEO of OpenAI, creator of ChatGPT , and Demis Hassabis, Founder of DeepMind, signed a statement calling on governments to make tackling AI a «global priority» on par with pandemics and nuclear war.

For the first time, the world's leading AI leaders have collectively called for urgent action to address the risks to humanity related to technology.

The Center for AI Security, a non-profit organization, hosted the statement, said it was designed to «raise concerns about some of the most serious risks of advanced AI» amid a «wide range» of concerns about the technology.

“Reducing the risk of extinction due to AI should be a global priority along with other societal risks such as pandemics and nuclear war,” said the statement, which was also signed by former Google AI chief Geoffrey Hinton and Dario Amodei, head of the Artificial Intelligence Lab. Anthropic.

0605 Artificial Intelligence Impact

Last week Mr. Altman, Mr. Hassabis and Mr. Amodeus met with Rishi Sunak to discuss risks including national security and «existential threats».

OpenAI, which has become the most famous AI developer in the world thanks to the rapid development of ChatGPT, has called for the creation of an international organization modeled on the International Atomic Energy Agency to manage this technology.

The short statement, signed by more than 350 policy makers and researchers, is meant to show a united front as politicians ponder how to regulate AI.

By comparison, the March letter signed by Elon Musk and others warned of a flood of propaganda and the call for a six-month pause in AI development divided the industry and swept through a string of fake signatures.

Dan Hendrix, executive director of the AI ​​Security Center, said the warning was «a situation reminiscent of nuclear scientists' warnings about the very technologies they created.”

Mr. Altman wrote last week that existential risk from AI could emerge within a decade as software outperforms human experts in most cases. fields.

However, he also suggested that OpenAI might respond to Brussels' new AI rules by shutting down ChatGPT in the EU. He later retracted the claim.