Technology company OpenAI is considering slowing down the development of its most advanced artificial intelligence (AI) models. According to anonymous sources from Bloomberg, the company’s CEO Sam Altman told employees at an all-hands meeting this week that he hopes other AI companies will do the same.
You might be interested: Were you lured with quick money? Pyramid scheme aka airplane stripped thousands of people of their savings
Concerns about autonomous agents growing. OpenAI slows development despite planned IPO
Recent weeks have seen incidents that have contributed to growing concerns about the risks of autonomous AI agents that can circumvent rules and exploit unexpected ways of interacting with external systems. Altman told employees that OpenAI is seeking coordination with other companies in the field, but some of them may not agree with the proposal.
In recent weeks, some employees at leading AI companies have publicly expressed concerns about the increased risks of advanced AI systems. According to Bloomberg, this has raised alarm within these companies. Both OpenAI and rival Anthropic filed confidential documents required for going public earlier this year.
OpenAI previously stated that it has recently slowed some parts of model development and paused some internal AI training due to safety concerns. In July, Altman said he had spoken with White House officials about the need to slow the pace of AI development.
In recent months, several employees from leading AI safety research labs have resigned. Anthropic researcher Jacob Coxon announced this week that he is leaving the company and the field due to concerns that Anthropic and its competitors are developing self-improving systems too quickly, which he believes could escape human control.
Don’t miss: TradeLocker vs MetaTrader
“Is the endgame approaching?” Researchers fleeing AI labs warn of loss of control
He fears that such systems could destroy humanity. According to him, many colleagues in the field are now using words like “crunchtime” (decisive moment) and “endgame” (final phase) when describing the direction of self-improving models, terms indicating that development is reaching its finale.
Coxon left OpenAI earlier this year and joined Anthropic because, according to him, this company is known for its efforts on model safety. Although he considered Anthropic’s safety efforts to be genuine, he now believes that no company can responsibly develop AI that would surpass humans in a range of tasks without government intervention or coordinated industry slowdown.
At the end of July, more than 1,000 employees from all major artificial intelligence companies signed a petition demanding the introduction of a mechanism to slow the pace of artificial intelligence development.
Read more: FX Junction
Source: ČTK
