AI · Security Desk · continually updated
AI security
The measured side of AI security: what the bug bounty platforms pay, which vulnerabilities AI tools have found, how each frontier model does on security work and what it will answer, and the incidents as they are logged, with the reporting beside the data.
- programmes measured
- 1,178
- AI-found CVEs at a record
- 47
- models scored
- 30
- incidents logged
- 92
- guides
- 5
The trackers
Each one downloads as CSV and JSON, CC BY 4.0Incident logAI Risk RadarThe standing log of AI-security incidents since 1 January 2026: prompt injection, hijacked agents, poisoned and leaked models and deepfake fraud, with the source on every record.92 incidents · 15 seen in the wildOpen AI Risk Radar Updated 18 Sept 2026Measured datasetBounty economicsWhat 1,178 public bug bounty programmes on seven platforms pay at their ceiling, how fast they answer, who gives safe harbour, and each programme's rule on AI-assisted reports.1,178 programmes · $4,000 median Bugcrowd ceilingOpen Bounty Economics Updated 19 Sept 2026Live registerAI-found vulnerabilitiesEvery vulnerability an AI tool found and a public record credits, each checked at its official CVE record or vendor advisory, with the strength of the evidence beside the credit.47 checked at a record · 69 entriesOpen the register Updated 19 Sept 2026Scored boardModel security capabilityHow each frontier model does on security work, with the harness, the task count and the safeguard state beside every score, and what each lab lets it answer.30 models · 50 scores · 18 independentOpen the board Updated 19 Sept 2026
