Google pays record $250,000 bounty for AI-assisted Chrome exploit affecting 1 billion users

Google has made a record payout of $250,000 to security firm XBOW after its team used a public AI model to build a Chrome exploit affecting roughly 1 billion users. It is the first time Google has paid its top full-chain reward under the updated 2026 Chromium VRP rules, and it took just six days.

The security firm had even teased the reward in a tweet in September:

For the most part, we’re told to be careful when clicking on links, buttons, or even downloading files on sketchy websites. However, as researchers at XBOW shared, this attack could have hijacked a Windows PC the moment someone loaded a malicious webpage without clicking anything.

Normally, Chrome traps webpage code inside two protective sandboxes so a shady site cannot touch the rest of your computer. To break through both barriers, researcher Brendan Dolan-Gavitt and AI agents running OpenAI’s public GPT-5.6 Sol model found and linked three separate bugs across Chrome’s JavaScript engine, WebP image handler, and crash reporter.

Running those AI agents over six days in August cost about $40,000 in OpenAI API charges, according to XBOW’s technical write-up. “Pre-LLMs a Chrome full chain would take a team of four around 6-9 months to be produced, our first submission took 2 months. The awarded one? six days,” Dolan-Gavitt wrote.

Their first attempt nearly fell apart when Google noted one of the bugs had already been reported, so the team swapped in a backup WebP flaw they had ready and kept the chain alive. Google quietly patched the flaws in late August and early September, starting with the Chrome 152 update we covered that squashed over 300 bugs.

Seeing AI models show up in Google’s bug reports is quickly becoming the norm, especially after the massive Chrome 155 update we reported just last week, where both OpenAI’s Codex team and Anthropic’s Claude were credited among 247 security fixes.

That surge in AI-assisted submissions is also forcing Google to rethink how it pays researchers. According to SecurityWeek’s report on Google’s Chrome bug bounty changes, the company just partially reversed its steep April payout cuts.

It’s now quadrupling base rewards for memory-safety flaws from $500 to $2,000 (with payouts reaching up to $20,000 for critical process bugs and sandbox escapes), even as it temporarily paused its Open Source Software Vulnerability Reward Program because of a flood of AI-generated reports.

We stand out from the tech-media crowd because we break news stories; we mainly bring you stuff that you won’t find anywhere in the mainstream tech media. Our stories have been picked up by some of the world’s most popular websites and media outlets—more info is available here.