Thursday, November 19, 2026 Updated Briefings Saved 中文
AI Daily Radar

AI and the frontier tech around it, explained every morning

Announced

Anthropic shortens vulnerability disclosure window to 30 days for severe bugs

Photo: cookieone / Pixabay

Anthropic said it is cutting its coordinated vulnerability disclosure window for high- and critical-severity issues from 90 days to 30 days, citing AI-driven bug discovery as the reason. The change was stated in an Anthropic research post, in which the company said it remains “optimistic that this will help defenders find and fix vulnerabilities faster than ever before.” The post, as relayed by Techmeme, does not give an effective date, does not say whether the shorter window applies to medium- and low-severity reports, and does not explain what happens if a fix is not ready when the 30 days expire.

Verified and unconfirmed

  • Verified: Anthropic’s own policy announcement says the coordinated vulnerability disclosure window is cut from 90 days to 30 days for high- and critical-severity issues, and set at 45 days for all other issues; Anthropic attributes the change to AI-driven vulnerability discovery and says it expects defenders to find and fix bugs faster (Anthropic, via tool-2-1).
  • Not yet confirmed: the effective date, whether the shorter deadline applies to bugs already in the disclosure pipeline, and any measure of how much faster AI tooling is actually surfacing flaws — the AI rationale and the “faster than ever before” expectation are Anthropic’s own claims, not independently tested results; some third-party write-ups still describe the 90-day window as Anthropic’s current standard (tool-2-2), so practical uptake of the new timeline is unverified.

Background

Coordinated vulnerability disclosure is the practice of giving a vendor private time to fix a reported security flaw before details become public. Anthropic had used a 90-day window for high- and critical-severity issues; it is now shortening that window to 30 days and says the change responds to AI-driven bug discovery, aiming to help defenders fix flaws faster.

Why it matters

Anyone who reports a high- or critical-severity flaw to Anthropic now has a 30-day clock instead of 90 before coordinated public disclosure, so researchers and downstream maintainers who rely on the longer window should plan for fixes to be ready in roughly a third of the previous time. Security teams that track Anthropic’s advisories should also expect fixes for serious issues to appear on a tighter schedule, since the policy is explicitly framed around accelerating defender response rather than around a model or product release.

What to watch

Watch for Anthropic to publish follow-up data on how many high- and critical-severity reports it actually resolves within the 30-day window, and for whether other AI labs with similar disclosure policies shorten their own deadlines.

Further reading

Comments

Comments

Comments use a GitHub account and are moderated. Comment rules

More in AI

See all