Google finds an AI that hunts and exploits software flaws without a human
Google has spotted an AI tool that finds and exploits software flaws entirely on its own, no human needed. The files inside it carry Russian-language comments, according to CoinDesk. It is a step beyond the AI assistants that still need someone at the keyboard.
- Original source: coindesk.com.
- The source published it on Nov 18, 2026 at 7:37 AM.
- No second outlet has reported it yet. Treat it as a single-source claim.
Google has found an AI tool that hunts down software flaws and exploits them on its own, with no human at the controls. The discovery came with a twist: the tool's files carry comments written in Russian, according to CoinDesk.
What sets it apart is the autonomy. This is not an assistant waiting on a person to pick a target and pull the trigger — it runs itself, spotting vulnerabilities and going after them without a hand on the wheel.
The find lands in the middle of a fast-moving debate over how far AI agents should be allowed to roam. A tool that clears the human out of the loop turns a flaw-finding model into something closer to a self-driving attacker.
TopicsGoogle
Source: coindesk.com
Written by AI from verified reporting.
More stories
- Tech & AI blockchain.news 1 viewXingchen unveils Xing 4, a 285B-parameter model built on Ascend chips
- Tech & AI x.com 1 viewCalifornia AG says his office is looking into the Nvidia scandal
- Tech & AI cryptobriefing.com 1 viewNvidia's Singapore Unit Books $104B for 2025 as Reporting Rules Shift
- Tech & AI x.com 1 viewNvidia Board Faction Pushes for Jensen Huang's Return as CEO, per Semafor

