Archive · 14 Sep 2026
Sam Altman urges stronger safeguards as AI capabilities rapidly advance
Sam Altman says artificial intelligence should always help people.
He worries that very powerful AI could become difficult for humans to control.
He also worries that one person, company, or country could use AI to force its ideas on everyone else.
Altman says safety work must keep up with the speed of AI development.
He believes companies should check for dangers while building AI, not only after it is finished.
He supports common safety rules and independent reviews.
He says companies should earn public trust without waiting for new laws.
He also says careful testing may take more time and money, but competition should not lead to reckless decisions.
OpenAI CEO Sam Altman warned that humans could lose control of increasingly powerful artificial intelligence systems.
Altman said AI must serve people and rejected allowing one person, company, or country to impose its worldview through AI.
He argued that safety measures must improve faster than AI capabilities and begin during development, not only before release.
Altman supported common safety rules, independent checks, shared safety methods, and stronger public transparency from leading AI companies.
He said slowing progress would not mean stopping development, but could require more time and money for testing, monitoring, and safety.
- Who
- Sam Altman, the chief executive of OpenAI, and leading artificial intelligence companies.
- What
- Altman called for stronger safety measures to prevent humans from losing control of powerful AI and to reduce the concentration of AI power.
- When
- In a post on X; the article does not provide a specific date.
- Where
- The comments were posted on X and discussed in relation to AI companies and the United States government.
- Why
- Altman said AI capabilities are advancing rapidly, creating risks that safety, monitoring, and public accountability must address.
- Speaker
- Sam Altman, OpenAI chief executive
- Primary concern
- Humans could lose control over powerful AI systems
- Second concern
- AI power could become concentrated in one person, company, or country
- Safety approach
- Safety arguments should be prepared before training steps expected to significantly increase capabilities
- Oversight
- Altman supports independent checks and common safety rules
- Industry responsibility
- Companies should build public trust without waiting for new laws
- Cost of caution
- Testing, monitoring, and safety may require additional time and money