OpenAI CEO Sam Altman has stated that society must be willing to accept some negative consequences of artificial intelligence in exchange for the benefits this technology brings.
In an interview with Politico, Altman also made clear the difference between OpenAI's approach and that of rival company Anthropic, arguing that AI should remain widely accessible and that its benefits can justify accepting some risks.
"We believe that the world should accept that some bad things will happen in exchange for the benefits that this technology brings and the opportunity for people to have control over their decisions," Altman said.
His stance comes at a time when the artificial intelligence industry is facing growing pressure to slow the development of increasingly powerful models and strengthen safety mechanisms.
At the center of the debate is the stance of Dario Amodei, the head of Anthropic, who has called for AI companies to slow down the pace of improving the capabilities of more advanced models, so that developers and independent evaluators have more time to identify and address risks. Altman has partially supported this idea, but it appears that the two companies continue to have significant differences on how the technology should be regulated.
Altman particularly opposed the idea that very powerful artificial intelligence should be controlled by a very limited number of companies or labs. According to him, such a concentration of control would be a “completely unacceptable compromise.”
The head of OpenAI argued that he would not accept an agreement that aimed to eliminate every risk, including hacking, misuse of technology, or fraud, just to guarantee development without negative incidents.
According to Altman, the benefits that people can gain from artificial intelligence will far outweigh the harms that can be caused by its misuse. He supports a lighter regulatory approach, emphasizing the importance of broad public access to this technology.
However, Altman set a clear line as to what risks he said were unacceptable. He said society should not tolerate “truly catastrophic” risks, especially a serious loss of human control over artificial intelligence systems.
The debate between OpenAI and Anthropic reflects one of the main issues in the AI industry: how quickly this technology should develop and to what extent society should accept its risks in exchange for potential benefits.
