Skip to content
NewsIncident

Google pauses open source bug bounty payouts after a flood of invalid AI reports

· by Pondero Newsdesk

The short version

Google stopped accepting new product-vulnerability reports to its Open Source Software Vulnerability Reward Program on October 1, 2026, citing a flood of invalid, AI-generated submissions that overwhelmed its triage team.

Google pauses open source bug bounty payouts after a flood of invalid AI reports

Google's own bug-hunting program just got buried by the same kind of automation it was built to reward. Since October 1, 2026, Google has stopped accepting product-vulnerability submissions to its Open Source Software Vulnerability Reward Program (OSS VRP), and it will not say when the door reopens.

What

Google's bug-bounty rules page now states plainly that "as of October 1, 2026, we are no longer accepting product vulnerabilities submitted to the OSS VRP," with a commitment to "give an update in Q1 2027" per Google's own program rules page. The pause covers only product-vulnerability reports, the category where researchers flag code defects, logic flaws, or design bugs in Google's public repositories. OSS VRP supply-chain reports keep running, and Google wrote that it may still accept some product-vulnerability reports through its separate Cloud VRP for Google Cloud repos. Reports filed before October 1 are unaffected.

Google told reporters the reason directly: "This pause is due to a significant rise in automated submissions, the vast majority of which are not valid," per TechCrunch. Google engineers and open-source maintainers had reportedly been spending hours manually validating reports that turned out to be hallucinated or unexploitable, time that came directly out of fixing real vulnerabilities, per Tom's Hardware. In the meantime, Google is pointing participants toward its other VRP tracks and the Patch Rewards Program instead.

The program's own reward table shows the gap in plain numbers. Supply-chain compromise reports, which are still being accepted, pay $500 to $31,337.70 depending on how critical the affected project is per Google's rules page. The product-vulnerabilities row of that same table now reads a flat dash across every project tier, meaning there is currently no dollar amount Google will pay for that category at all, not a reduced one.

A paying channel for bug hunters just closed until 2027

For any researcher who relies on AI tools to scan Google's open-source repositories for flaws, the biggest named payout channel for that work is gone for at least a quarter, with no guaranteed return date. Submitting a low-effort, AI-generated report to OSS VRP no longer has a payoff, which should push serious hunters back toward Google's other bounty programs, the Cloud VRP for in-scope Cloud repos, or toward manual verification before they file anywhere else. Anyone still running a bulk AI-scanning workflow against Google's public repositories and expecting a check for each hit now has one fewer place to send the output.

The deeper signal is for anyone building or selling AI-assisted security-research tools. A program that once rewarded speed and volume is now the one punishing it, and the fix Google reached for was not a lower reward tier for AI-assisted reports; it was zero reports accepted at all. That is a harder line than most vendors have drawn so far, and it raises the open question of whether gating out automation entirely, rather than just screening it more carefully, becomes the standard response once report volume outpaces triage capacity.

Context and reactions

Google's move lands in a year when AI-written reports have strained several major open-source security efforts at once. Linux kernel maintainers said this month they were "completely overwhelmed" by CVE submissions as the kernel approached a record 2,000 vulnerabilities per release, with AI bug hunters scouring roughly 40 million lines of code, per Tom's Hardware's separate report. Intel suspended its own bug bounty program, which had paid up to $100,000 per flaw, and replaced it with a no-reward disclosure process through Intigriti; Intel did not confirm AI-generated reports caused the move, but security researchers Tom's Hardware spoke with suspect that is exactly what happened, per Tom's Hardware. TechCrunch had flagged the underlying risk more than a year earlier, reporting in July 2025 that cybersecurity experts were already warning that AI-generated slop posed a serious threat to bug bounty programs industry-wide, per TechCrunch's 2025 report.

What ties the Google, Linux, and Intel cases together is cost asymmetry. Writing a plausible-sounding vulnerability report with an AI agent now takes minutes and costs almost nothing, while verifying whether that report describes a real, exploitable flaw still takes a trained engineer reading actual code. Google, Intel, and the Linux kernel team each hit that same wall from a different angle this year, and each one responded by cutting off the submission channel rather than trying to out-hire the flood.

What to watch next

Google's promised Q1 2027 update will show whether OSS VRP reopens with new friction built in, such as requiring a reproducible exploit or OSS-Fuzz confirmation before a report qualifies, or whether product-vulnerability submissions stay closed indefinitely. Also worth tracking: whether third-party platforms like HackerOne and Bugcrowd adopt similar AI-submission screening, especially after Intel already dropped per-flaw cash rewards entirely in its own program redesign.

Sources