Sam Altman warns humans could lose control of AI

OpenAI CEO Sam Altman warned in recent public statements and interviews that humans could lose control of advanced artificial intelligence as models grow more capable.

He cautioned that advanced systems may behave in ways their creators do not expect or cannot manage, and that systems could pursue objectives misaligned with human intentions.

Altman urged increased investment in alignment research, which aims to make AI intent and behavior match human goals, and called for coordination among governments, companies and researchers on rules and testing before more powerful systems are released.

He set out practical measures to reduce risk, including stricter pre-deployment testing of new models, independent audits of capabilities and safety, clearer reporting requirements for developers, and stronger government standards and enforcement to supplement company efforts.

OpenAI has expanded internal safety work, including teams focused on alignment and red-teaming to probe models for harmful behavior. Altman said industry-wide approaches are needed because a single organization may not anticipate every danger and because shared research can reveal unexpected system capabilities.

Observers in the technology sector noted Altman’s remarks reflect ongoing debates over the pace of AI development and the balance between innovation and risk management. Lawmakers in several countries have held hearings and are considering legislation that would require transparency, risk assessments and controls for powerful models.

Critics have pushed for clearer timelines and measurable safety benchmarks. Altman acknowledged those critiques and called for operational rules and concrete tests to determine when a model is safe to release.

Concerns about losing control of AI have circulated among researchers and policymakers for years and intensified as generative models showed unexpected abilities in language, code and decision-making tasks. The field of alignment develops methods designed to ensure AI systems follow human instructions and respect human values.