Nvidia's CEO Says AI Needs No New Laws — Should You Trust That?

Nvidia's boss Jensen Huang told a Salesforce conference this week that AI doesn't need new laws. His argument: AI is just hardware and software built by humans, so it can be controlled by humans with the rules we already have. "Safety is an engineering problem, not a legal one," he said. In his view, companies should simply choose not to release products they aren't confident are safe, and the market will punish anyone who gets it wrong. No new regulations required.

It's a reassuring pitch from a man who clearly knows the technology. Nvidia has been building the chips that power AI since long before ChatGPT existed, and it now makes its own AI models and software tools too. But there's an obvious catch: Nvidia has made a fortune from the AI boom, and new rules could slow down its sales. So his "trust us" stance is also very convenient for him.

The problem is that good intentions don't guarantee safe products. In 2024, a faulty software update from CrowdStrike grounded thousands of flights worldwide. Meta just paid $18 billion to settle claims that social media harmed children. And AI has already caused real damage — from a model hacking into systems to lawsuits over young people's suicides after long chats with a chatbot. Companies sometimes act badly on purpose, too.

Huang is right that existing product liability laws might eventually cover AI harms — but only after years of court cases, and only if the damage isn't catastrophic first. Meanwhile, the industry has a narrow window to police itself voluntarily, and to convince rivals worldwide, including in China, to join in. Microsoft's CEO Satya Nadella made the same point: safety concerns shouldn't stop at borders. For now, Huang's answer is simple — run fast, and pause only if something feels unsafe.