Nvidia CEO argues AI safety is an engineering problem, opposes new regulation

Nvidia founder and CEO Jensen Huang argued at Salesforce’s Dreamforce conference that AI safety is an engineering problem, not a legal one, and therefore does not require new regulation. He described AI as fundamentally hardware and software built by humans, implying it can be controlled through existing laws and corporate responsibility.

Huang stated that companies should only release products they are confident are safe, and that market forces will pressure firms to avoid releasing unsafe AI systems. He framed innovation, speed, and safety as compatible goals, urging companies to "run as fast as you can" while pausing when safety is uncertain.

He also positioned open-weight models as a competitive counterweight to proprietary AI labs, but did not endorse formal industry self-regulation or new legislation. Huang’s remarks come amid ongoing debates about whether AI-specific laws are needed to address risks such as misuse, unintended harms, or systemic failures.

The comments follow incidents cited by critics as evidence that current safeguards are insufficient, including a 2024 CrowdStrike outage that disrupted global operations and lawsuits alleging harms linked to AI systems. Huang did not address whether existing product liability frameworks are adequate for AI-related damages or whether new legal precedents are required.