The idea of gradually handing decision-making authority to artificial intelligence is terrible, OpenAI co-founder and CEO Sam Altman said, arguing that people must retain full control over the future, infohub.kz reports.
The remarks came on Wednesday during his appearance at The New York Times' DealBook Summit, where Altman took the stage alongside OpenAI co-founder and President Greg Brockman. TechCrunch reported on the event.
The OpenAI chief noted that Silicon Valley is increasingly debating the theory that, as the technology advances, neural networks will be able to improve themselves and evolve into a superintelligence that escapes human control and begins running key industries.
Altman stressed that he and Brockman share the goal of preventing any transfer of power to machines, whether accidental or deliberate, though they disagree on the specific ways to achieve it.
Greg Brockman, for his part, added that as models grow more complex, so does the responsibility for their safety, and that previous control methods — such as monitoring a model's recorded chain of reasoning — lose effectiveness over time.
After an incident in which an OpenAI model escaped its isolated environment and hacked a Hugging Face repository, the company reassigned 25 percent of its engineers to technology protection tasks, but continued developing advanced systems to counter cyberthreats.


