Only 12% of Top AI Firms Run Bug Bounty Programs

A Cybernews ranking of the top 500 AI companies finds most lack publicly identifiable programs for reporting security flaws.

Updated on Nov. 17, 2026 in Artificial Intelligence

Only 12% of Top AI Firms Run Bug Bounty Programs

Live Poll

Do you trust AI companies to keep your personal data secure?

The Cybernews AI Trustworthiness Ranking found that 62 of the top 500 AI companies have a public bug bounty or coordinated vulnerability disclosure program, while 438 have none identified. OpenAI has raised rewards in its Security Bug Bounty Program, with payouts of up to $30,000 for critical ChatGPT, Codex or API Platform vulnerabilities and up to $172,000 for exceptional findings with bonuses.

Why it matters

Public bug bounty and disclosure programs let security researchers report flaws before malicious actors exploit them. The ranking counts only publicly identifiable programs, so internal security teams, penetration tests, audits and private invite-only bounties are not reflected in the figures.

The Cybernews ranking reflects only publicly identifiable programs, with 62 of the top 500 AI companies (12%) listing one and 438 showing none. Internal teams, audits and private bounties are excluded, so the full extent of vulnerability testing remains unmeasured.

The players

Cybernews

Cybernews is a cybersecurity news and research outlet that publishes independent analysis of online threats and digital risk.

OpenAI

OpenAI is an artificial intelligence research company that developed ChatGPT and operates the Codex and API Platform products.

The details

Google, Microsoft, Meta, Anthropic, Mistral AI, Perplexity AI and Hugging Face are among the companies with bug bounty or CVD programs, while Midjourney, ElevenLabs, DeepSeek, Cursor, Runway and Suno have none identified. OpenAI's expanded program covers ChatGPT, Codex and its API Platform.

Timeline

  1. Cybernews published its article on AI bug bounty programs on November 17, 2026.

The Tech Race

Public bug bounty and coordinated vulnerability disclosure programs, which software makers adopted over the past two decades, are now being tested against the fast-growing AI sector. The ranking shows that large AI firms have been slower to adopt the same public disclosure channels that established software companies such as Google and Microsoft already run.

Users of ChatGPT, Codex and other AI tools rely on companies to fix security flaws quickly, and public programs are one mechanism for that. Users of AI products from companies without identified programs have fewer publicly documented channels for reporting vulnerabilities.

The takeaway

Checking whether an AI provider publishes a vulnerability reporting program is a simple way to gauge how seriously it handles security. Programs that offer public rewards signal a formal process for flaw disclosure.

Further reading

For more on how AI companies handle security and model risks, see our Artificial Intelligence coverage.

Source note: This article includes information reported by Cybernews.

Live Poll

Do you trust AI companies to keep your personal data secure?