Sam Altman Calls for Transparent Safety Standards in AI

Sam Altman, CEO of OpenAI, told attendees at Salesforce's Dreamforce conference that certain incidents involving AI technologies are inevitable as the field advances, and stressed the need for a transparent reporting and learning process.

Learning from aviation safety

Altman highlighted the success of bodies such as the U.S. Federal Aviation Administration (FAA) and the National Transportation Safety Board (NTSB) in improving safety through continuous incident reporting and investigations, urging the AI sector to adopt a similar approach.

He said AI companies must be ready to implement “slow‑down or halt” measures when safety risks emerge, placing ethical standards above the race for technical prowess.

Industry backdrop

The remarks come amid internal warnings from former OpenAI and Anthropic researchers about the dangers of self‑improving AI systems that could escape human control, fueling debate over how fast development should proceed.

Altman emphasized that the goal is not to accept crises but to detect them early, learn quickly, and deploy effective safeguards to prevent escalation.

Through this framework, the tech community aims to create a safer AI environment that sustains innovation while mitigating potential hazards.