Google's New AI Can Find Software Flaws Before Hackers Do

Google just unveiled Gemini 4 Carbon, its most advanced AI yet, with a big focus on cybersecurity. This AI is really good at spotting weaknesses in software—the kind hackers exploit—and at writing code that's secure from the start. It's like having a super-smart security guard that never sleeps, constantly checking for open windows and locks. For now, only a few cybersecurity companies can use it, but soon more partners will get access.

Why should you care? Every time you use an app, shop online, or share data, you're relying on software being secure. Data breaches cost companies billions and expose your personal info. An AI that finds bugs faster means fixes happen sooner, reducing the chance your data gets stolen. Google says Carbon can even work on complex tasks for days, tracking down tricky problems that humans might miss.

But there's a catch: the same AI could potentially be misused by hackers to find vulnerabilities first. Google is limiting access to trusted defenders for now, but it's a double-edged sword. Also, Carbon is pricey—$2 per million input tokens and $10 per million output tokens—though that's competitive for such advanced AI. For now, it's a tool for the pros, but its impact on your digital safety could be huge.