AI-Generated Slop Floods Bug Bounty Programs With False Reports
HackerOne and open-source maintainers report a surge in AI-generated vulnerability reports that overwhelm security teams with technically plausible but incorrect data.
The proliferation of AI-generated content, termed slop, is transforming bug bounty programs into security liabilities. HackerOne reported a surge of over 100% in report volume following the release of advanced AI models in February 2026, forcing security experts to spend critical hours debunking polished but incorrect reports instead of remediating actual risks.
This trend has severely impacted open-source projects. Daniel Stenberg, a maintainer of the tool curl, eliminated monetary rewards for the project in January 2026 after the rate of confirmed vulnerabilities plummeted from 15% to below 5%. While the quality of reports later recovered to 15-16% after curl returned to HackerOne without financial incentives, the overall volume remained double the 2025 rate.
To provide maintainer relief, the curl project stopped accepting reports entirely in July 2026. Ido Geffen, CEO of Novee Security, analyzed the impact of this trend, noting that the technical plausibility of AI-generated noise makes it increasingly difficult for teams to filter out false positives.