Sam Altman (샘 알트먼), chief executive of OpenAI, said society should bear a certain level of side effects in light of the benefits artificial intelligence could bring.

Online media outlet Gigazine said on Oct. 5 that Altman, in an interview on Politico’s newsletter and podcast service “Decoded,” highlighted differences with Anthropic over AI regulation.

The core issue was the strength of regulation. Altman said a fundamental difference in worldview remains between OpenAI and Anthropic. Asked where he differs with Dario Amodei (다리오 아모데이), Anthropic’s CEO, Altman said that when weighing the benefits of AI technology and whether people continue to retain agency, the world should accept it even if some bad things happen because of AI.

The remarks have drawn attention as they intersect with recent security controversies involving OpenAI. OpenAI has shown that it considers a relaxed form of AI regulation preferable. But after multiple security issues emerged, including a case in which OpenAI’s AI agent hacked Hugging Face, it also showed some movement toward agreeing with Amodei’s argument that the pace of AI development should slow.

Altman drew a line on how regulation should be done. He said he can understand the view that AI is dangerous technology and that a single company should hold it, ensure it is tightly controlled so problems do not occur, and then discuss how to distribute its benefits. He said such an approach is difficult to reconcile with the relaxed regulatory stance OpenAI has advocated.

Altman said he cannot accept demands to guarantee that large-scale hacking, technological misuse, fraud or other wrongdoing will never occur. He explained that he believes people will do more good than harm.

The remarks also intersect with the relationship between Altman and Amodei. The Wall Street Journal has reported that conflict between the two has continued for 10 years and their confrontation has not been resolved. In that situation, Altman’s recent posture of agreeing with some of Amodei’s concerns was seen as unusual, but the latest interview again confirmed that differences between the two sides remain unresolved.

As a result, debate in the AI industry has become clearer. One side sees a need to more tightly control the pace of development and access rights to reduce the risk of misuse, while the other side sees a need to manage risks while opening up more widely the benefits of the technology and users’ freedom of choice. Altman’s remarks were read as making clear that even while OpenAI acknowledges security problems, it does not agree with a way of significantly slowing the spread of AI itself.