Sam Altman Urges Accepting Risks for AI Benefits
The head of OpenAI told Politico in an interview that restricting access to artificial intelligence could do more harm than good.
Sam Altman believes the world must be prepared to accept some negative consequences of artificial intelligence development to avoid depriving people of its benefits. According to him, attempts to completely eliminate risks, such as hacking attacks or fraud, could lead to excessive restrictions on the technology and the concentration of control over it in the hands of a small number of companies, individuals, or a single country.
Altman emphasized that regulation should focus on the most serious risks, including the possibility of people losing control over AI systems. At the same time, he acknowledged that it is impossible to completely prevent the misuse of the technology. In August, OpenAI advocated for strengthening California’s SB 53 law, which requires developers of advanced AI models to disclose their safety approaches and report certain critical incidents to authorities.
This year, during internal cybersecurity tests, OpenAI’s models exceeded the intended boundaries of the test environment and gained unauthorized access to Hugging Face’s systems. In a separate incident, during the training and testing of the company’s model, unauthorized access was made to websites of Australian government agencies. Anthropic, a competitor of OpenAI, also reported four incidents in which Claude models, during cybersecurity tests, gained unauthorized access to real systems of third-party organizations.
