New York has taken a bold leap into AI safety by passing the RAISE Act, aiming to regulate frontier AI models and prevent disasters involving massive casualties or damages. A win for transparency and accountability amidst industry pushbacks.

📰 New York Passes Ambitious AI Safety Bill to Prevent Technological Disasters

Hold on to your hats, AI enthusiasts and concerned citizens alike! New York just became the first state to pass the RAISE Act, a bill that aims to slap some seriously needed regulation on frontier AI giants—think OpenAI, Google, and Anthropic—to prevent AI-fueled disasters. We're talking scenarios that could lead to the injury or death of over 100 people or damages north of $1 billion. Yup, they mean serious business.

💡 What’s the RAISE Act All About?

This isn't your run-of-the-mill policy fluff. The RAISE Act aims to:

Demand transparency: AI labs building cutting-edge models must disclose detailed safety and security reports.
Incident reporting: Any safety mishaps or security breaches, like rogue AI behaviors or stolen models, have to be reported.
Enforce consequences: New York's attorney general can slap civil penalties up to $30 million if companies don't get with the program.

Fun fact: The bill targets AI players whose models have been trained using more than $100 million in computing resources and whose products are available to New Yorkers. So startups and small-time researchers can breathe a little easier.

📜 A Win For the AI Safety Movement

The RAISE Act signals a huge win for the AI safety squad—think Nobel laureates like Geoffrey Hinton and pioneers like Yoshua Bengio—who want to see clear guardrails in a tech landscape sometimes too hungry for speed.

Why now? Senator Andrew Gounardes (co-sponsor of the bill) puts it bluntly: "The window to put in place guardrails is rapidly shrinking given how fast this technology is evolving." He warns these risks are "incredibly likely," which isn't exactly the bedtime story you want.

⚖️ Designed to Protect Innovation Too

If you’re worried this is just another Silicon Valley killing machine, fear not! The RAISE Act was carefully crafted not to suffocate startups or academic researchers. That's a direct response to California's earlier SB 1047 AI bill, which was famously vetoed for being too harsh and innovation-crushing.

🙅‍♂️ Industry Pushback & Critiques

Surprise, surprise—Silicon Valley isn't throwing a parade for the RAISE Act. Andreessen Horowitz’s Anjney Midha called it "yet another stupid, stupid state level AI bill..." (yeah, not holding back).

Anthropic’s co-founder Jack Clark expressed concerns about the bill's scope, especially how it might inadvertently strain smaller AI companies. Meanwhile, OpenAI, Google, and Meta have been quiet, at least publicly.

🤔 What’s Not in the Bill?

Notably absent? The dreaded "kill switch" compulsion on AI models and accountability for companies if their post-trained frontier AI models cause critical harms. The bill opts for a lighter touch on some of these points.

📍 Why Does This Matter? And Who Cares?

For AI users and enthusiasts: This bill could make AI safer and more transparent.
For startups and researchers: The bill eases fears of overbearing regulation.
For policymakers: It’s an experimental blueprint that other states (or the feds?) might copy.
For the general public: It’s about keeping tech giants honest and protecting YOU from AI risks.

🔮 What’s Next?

The RAISE Act is currently sitting on Governor Kathy Hochul’s desk. She can sign it into law, send it back for tweaks, or veto it. Fingers crossed for the first option!

💬 Commentary:

So, New York just raised the stakes in AI safety law. It’s a mix of caution and pragmatism — acknowledging the lightning speed of AI innovation while trying not to rain on the innovation party too hard. Whether this bill will actually keep AI disasters at bay or ruffle the feathers of tech giants remains to be seen, but hey, progress is progress.

For those watching the AI legal landscape, the RAISE Act is definitely one to keep an eye on. It’s like the first chapter in a book about how society wrestles with powerful, world-altering tech—no spoilers, but it’s a thriller.

Stay tuned for more AI policy news and keep those algorithms friendly!

Image credits: Lev Radin/Pacific Press/LightRocket / Getty Images