Who Will Keep AI Safe? California Shows the Way

You've probably heard a lot about AI lately – those clever computer programs that can write, create art, and even help with customer service. As these tools get more powerful, people are starting to ask important questions about their safety. What if an AI makes a big mistake? Or, worse, what if it's used in a way that's harmful?

That's where California comes in. This US state, often a leader in new laws for technology, has just passed a new rule called SB 813. It says that by early 2028, independent organisations will need to check and approve the most advanced AI systems before they can be widely used. Think of it like getting a safety inspection for a new car model before it's allowed on the roads. The idea is to make sure these powerful AIs are tested for things like accuracy, fairness, and potential risks.

This move by California is a big deal because it’s one of the first times a government has tried to formally put a safety net around cutting-edge AI. It suggests that, just like medicines or planes, AI might soon need official 'tick of approval' before it reaches our homes and workplaces. It's a sign that governments around the world, including potentially Australia, are starting to think seriously about how to manage these rapidly evolving technologies.

One interesting point from the original report is the cost involved in these checks. A recent investigation into an AI incident cost around $400,000 in ‘tokens’ – essentially the computer power and resources needed to run these complex checks. This shows that making sure AI is safe isn't cheap or easy. It's a complex, technical job that requires a lot of expertise and resources. For Aussie businesses thinking about using AI, this also highlights the importance of choosing reputable AI providers who are committed to safety and ethical use.

Ultimately, this isn't just about big tech companies; it's about building trust in AI for everyone. If we know that the AI tools we use have been properly checked, we can feel more confident about their safety and reliability. This push for independent verification is a crucial step towards ensuring AI benefits society without introducing unexpected risks, making it safer for our kids, our businesses, and our daily lives.