The researcher wrote about this in his own post on the OpenAI blog.
He points out that, within a relatively short period of time, artificial intelligence models will be able to control computers, interact with people and other artificial intelligence systems, and carry out research projects. According to him, they will soon also be able to improve themselves without human intervention.
“These are times that call for extreme caution. I am concerned that no one is prepared for the consequences of the rapid continued growth of artificial intelligence,” wrote Pachocki.
The researcher writes that developers can either try to steer AI more closely towards human interests, or "slow down future development, if necessary". Pachocki expects that voluntary deceleration will become commonplace until "shared safety margins" are established.
Pachocki’s statement echoes the statements and predictions made by company executives and artificial intelligence developers. In May 2025, Dario Amodei, CEO of Anthropic PBC, warned that artificial intelligence could replace half of all entry-level jobs.
OpenAI CEO Sam Altman made a similar prediction back in February 2019, but in September 2026, in an interview with Bloomberg TV, he stated that the artificial intelligence industry is failing to effectively communicate the benefits of the technology to the public.
At the end of the summer, Microsoft co-founder Bill Gates published an essay on the implications of the rapid development of artificial intelligence. He called on governments to start preparing now for changes in the labour market, security risks and the impact of AI on human relationships.
Pachocki’s employer, OpenAI, has stated that it aims to create a "true automated AI researcher" in less than two years.
An article written by Yurii Svyrydiuk (Suspilne), initially published on 9 September 2026, 13:01 (CEST)