Photo by Yarenci Hdz on Unsplash
This week, California made headlines by signing a major artificial intelligence law. But here’s the twist: it’s the version Big Tech was rooting for.
On Monday, Governor Gavin Newsom officially signed the “Transparency in Frontier Artificial Intelligence Act” into law. Sounds like a big step, right? It kind of is. But depending on who you ask, it’s also kind of… not.
What the Law Actually Does
If you’re expecting hardline AI regulations with tightly enforced safety measures, you might be surprised. Instead of strict mandates, the new law is mostly about disclosure:
- AI companies making $500 million or more in annual revenue now have to publish their safety protocols online.
- They must report “potential critical safety incidents” to California’s Office of Emergency Services.
- Employees who flag safety issues will now have whistleblower protections.
- The California Attorney General can fine companies up to $1 million per violation if they don’t play by these rules.
The law also sets up something called CalCompute — a public computing cluster framework under a new government consortium — and the Department of Technology will recommend annual updates (though they’re just recommendations, not required changes).
So yes, there are some new obligations for companies. But here’s the catch: there’s no requirement for actual safety testing or independent oversight.
Photo by Clarissa Watson on Unsplash
What Got Left Out — And Why
This new law, known as S.B. 53, replaces an earlier proposal called S.B. 1047. That original bill, championed by Senator Scott Wiener, would’ve required:
- Mandatory safety testing for powerful AI systems
- Built-in “kill switches” to shut those systems down if things go wrong
It was intense. Critics even said it leaned too hard on sci-fi-style doomsday predictions. And Big Tech? They weren’t fans. They lobbied hard against it, calling the mandates vague and burdensome.
So last year, Newsom vetoed S.B. 1047. And the tech industry breathed a deep, well-funded sigh of relief.
This year’s S.B. 53 — the version that just passed — is what came after that. It focuses on transparency instead of strict controls. No testing mandates. No kill switches. Just talk-about-it public standards, many of which large AI companies already use.
Who Helped Shape the Law?
This updated bill didn’t come out of nowhere. It was shaped with input from AI experts like Stanford’s Fei-Fei Li and former California Supreme Court Justice Mariano-Florentino Cuéllar.
Even companies seem pretty happy with where things landed. Jack Clark, co-founder of Anthropic, called the law’s safety guardrails “practical.” Senator Wiener described it as “commonsense.”
That said, those “commonsense guardrails” mostly mirror existing internal practices used by big players. The law doesn’t define clear safety standards. And there’s no independent verification to make sure the disclosed safety practices are actually effective.
Why This Matters Beyond California
California isn’t just any state. It’s the epicenter of the AI world. The Bay Area alone drew more than half of all global venture capital funding for AI and machine learning startups last year. And according to the state government, 32 of the world’s top 50 AI companies are based there.
So what happens in California doesn’t stay in California.
Whatever laws it passes — even if they’re criticized for being soft — set a tone. They create precedents. They shape how AI safety is talked about at a national and global scale.
Right now, the signal seems to be that voluntary transparency is better than enforced compliance, at least when corporate influence is strong and the stakes are high.
Photo by Michelle Auger on Unsplash
Final Thoughts
This new AI law feels like a compromise — one that leans heavily in favor of the companies building powerful AI systems. It gives the appearance of responsibility and oversight without adding much real friction to tech development.
To be fair, that’s probably why it passed.
But if you’re hoping for more robust checks on AI safety moving forward, you’ll probably have to wait — or push — for more than just transparency.
And that’s the real takeaway here: this law tells us more about the current power dynamics between tech and policy than it does about actual AI risk prevention.
For now, California is asking Big Tech to be open about its practices — but not necessarily to change them.
Photo by Igor Shalyminov on Unsplash
Keywords: California AI Law, Artificial Intelligence, Transparency in AI, Big Tech, AI Regulations, Gavin Newsom, AI Safety.
