Sam Altman warns of rapid acceleration of AI before UN Security Council

The current progression of artificial intelligence (AI) “requires extreme attention,” warned OpenAI boss Sam Altman on Wednesday during a presentation to the UN Security Council. “As the (AI development) process becomes increasingly automated,” he said, “the pace of AI improvement could accelerate rapidly.”

Sam Altman was invited, along with his counterparts Dario Amodei, from Anthropic, and Frenchman Clément Delangue, from the AI ​​platform Hugging Face, to take stock of this technology and its risks. The Security Council meeting, initiated by France, takes place in a context of growing concerns around the trajectory of artificial intelligence.

Since the beginning of the summer, OpenAI and Anthropic have reported several incidents related to their most advanced models, which deviated from the objectives set by their developers, cheated, sought to lie and conceal their actions. The most serious slip-up saw two OpenAI AIs leave their confined environment during tests to go on the internet and intrude on Hugging Face, a platform for storing artificial intelligence models.

“We will do our best”

Furthermore, several professionals in the sector have warned of the threat that, according to them, AI poses to the future of humanity. At the beginning of September, a resigned Anthropic employee, previously with OpenAI, affirmed that, by their creators' own admission, this technology “could kill us all by the end of the decade”. “We do not want to fall into the trap of blind optimism,” argued Sam Altman, “but we cannot give in to catastrophism either. (...) So we will do our best to find a happy medium.

As he has already done on numerous occasions, the boss of OpenAI called for international cooperation on AI. “In history,” he recalled, “there have been times when countries which were in competition and did not necessarily like each other (...) still managed to agree for the common good in the face of a new technology. This needs to happen again.”