JupiteX Get the app
Science & Technology18 Nov 2026 · about 7 min

Google Project Zero cuts bug disclosures to 30 days amid AI threats

The brief

Google Project Zero is moving away from its “90+30” disclosure policy toward shorter, fixed deadlines. The change reflects the team’s view that AI has accelerated both vulnerability discovery and exploitation. Shorter deadlines aim to push vendors to investigate flaws and deliver fixes more quickly, but they also leave less time to prepare updates before details become public. The transition began on November 13, when Project Zero removed the 30-day extension from its timeline. The next specified step is February 2, 2027, when the deadline will fall to 60 days. The final scheduled reduction comes on January 3, 2028, when the deadline will become 30 days. The schedule is gradual rather than an immediate move to 30 days. Project Zero said this gives vendors time to improve their systems for managing and fixing flaws. Once the applicable deadline expires, the team will publish details even if a fix is not ready. The article does not state the year for the November 13 start date.

01

How quickly will Project Zero’s disclosure deadline shrink, and when will each change take effect?

Google Project Zero is moving away from its “90+30” disclosure policy toward shorter, fixed deadlines. The change reflects the team’s view that AI has accelerated both vulnerability discovery and exploitation. Shorter deadlines aim to push vendors to investigate flaws and deliver fixes more quickly, but they also leave less time to prepare updates before details become public.

The transition began on November 13, when Project Zero removed the 30-day extension from its timeline. The next specified step is February 2, 2027, when the deadline will fall to 60 days. The final scheduled reduction comes on January 3, 2028, when the deadline will become 30 days.

The schedule is gradual rather than an immediate move to 30 days. Project Zero said this gives vendors time to improve their systems for managing and fixing flaws. Once the applicable deadline expires, the team will publish details even if a fix is not ready. The article does not state the year for the November 13 start date.

02

What is Google Project Zero, and what does its vulnerability research team do?

Google Project Zero is a vulnerability research team focused on finding serious security flaws in widely used software and hardware. Its work matters because flaws in common products can leave users exposed, while the team’s reports can prompt vendors to investigate and prepare security updates. Its disclosure policy sets how long vendors typically have before technical details are published.

The article points to several recent findings: a .NET Framework zero-day, flaws in Chrome’s JavaScript engine and graphics components, and a Qualcomm chipset zero-day linked to NoviSpy spyware attacks. When Project Zero finds a flaw, it reports it privately to the vendor. That gives the vendor an opportunity to develop a patch before publication.

Project Zero typically publishes technical details after a set period, intended to allow time for fixes and for most users to apply them. The team is now shortening that period because it says AI has accelerated vulnerability discovery and exploitation. Under the planned schedule, vendors will face progressively shorter deadlines, and details will be published when the deadline expires even if a fix is unavailable.

03

What is vulnerability disclosure, and why do researchers usually give software vendors time to fix a flaw before publishing details?

Vulnerability disclosure is the process of informing a software or hardware vendor about a security flaw and, later, making information about that flaw public. Project Zero typically reports a finding privately first. The timing matters because public technical details can help people understand how a flaw works, while a vendor may still need time to prepare a response.

The usual delay gives vendors time to develop security updates and patches. It also gives most users a margin to apply those updates before the flaw’s details are released. For example, after Project Zero reports a flaw, it generally waits for a set period before publishing its technical details. That period is the disclosure timeline at the center of the policy change.

The balance is becoming harder, according to Project Zero. The team says AI has raised the chance that attackers will discover a flaw during an embargo and made it easier to reverse-engineer fixes once patches ship. Project Zero is therefore shortening its deadlines, while retaining a gradual transition and some exceptions. When a deadline expires, it will publish details even if no fix is available.

04

What can happen to users when a vulnerability’s details are published before a fix is ready?

If technical details become public before a fix is ready, users can face risk because the flaw remains unpatched while attackers can learn about it. The article highlights this concern in explaining why vendors opposed an immediate move to a 30-day deadline. Google said making unfixed vulnerabilities available during vendors’ remediation work could harm users.

Project Zero’s policy sets a point when details will be published, even if the vendor has not finished a fix. That can put pressure on vendors to investigate reports and ship updates faster. The article also says attackers are increasingly likely to discover vulnerabilities during an embargo and can reverse-engineer fixes after patches are released, reducing the security benefit of delaying disclosure.

The new schedule is phased in rather than changing immediately to 30 days. Project Zero said that approach gives vendors time to improve their flaw-management and remediation systems. Still, once the applicable deadline expires, publication can happen without a fix. The article does not describe specific attacks resulting from a particular Project Zero disclosure, but it notes vendors’ warning that premature disclosure could expose users to severe risks.

05

How can AI help attackers find vulnerabilities or turn a newly released patch into a working exploit?

Project Zero says AI has changed the pace of vulnerability research and attacks. It has accelerated the discovery and exploitation of vulnerabilities, the team says. That matters for disclosure timelines because a delay that once gave vendors and users useful preparation time may now also give attackers an opportunity to find a flaw or act on it.

The article describes one key mechanism: after a vendor releases a patch, attackers can reverse-engineer the fix to understand the vulnerability it addresses and develop a working exploit. Project Zero says AI has made this process easier. The article does not explain which AI tools are used or provide technical steps for carrying out that analysis, so it does not support more specific claims about how attackers do it.

Project Zero argues that the growing likelihood of attacker discovery during an embargo reduces the benefits of delaying disclosure. Its response is to shorten deadlines gradually, placing more pressure on vendors to investigate reports and ship fixes quickly. Details will be published after the deadline even if a fix is not available, under the new policy’s general rule.

06

When might Project Zero disclose a flaw early or hold back its details for longer?

Project Zero’s standard approach is to report a flaw privately and publish technical details after a set period. The article describes exceptions in both directions: the team may release information earlier in certain circumstances, or keep details back longer. These exceptions mean the deadline is a general policy, not an absolute rule for every reported vulnerability.

Early disclosure may happen if a flaw is actively exploited in the wild, if others have already published exploitation or vulnerability details, or if the vendor requests early disclosure. In those cases, the circumstances differ from a normal private reporting period, where the vendor is working toward a fix before details become public.

Google may also choose non-disclosure, as it deems fit. Project Zero gives Spectre and Meltdown as examples of cases where it withheld information for longer than 120 days. The article does not set out a complete test for when Google will use this exception. Its new shorter deadlines are therefore paired with discretion to handle exceptional cases differently.

07

How do software flaws arise, and how do patches repair them?

The article focuses on disclosure policy, not on how software flaws originate or how patches are built. As general background, software flaws can arise from mistakes or weaknesses in a product’s design, code, or handling of data. Such a flaw may create a security problem if it allows software to behave in a way its makers did not intend. This explanation goes beyond the article’s specific details.

A patch is a software update intended to correct a problem. In general, developers investigate a reported flaw, change the relevant software, and test the update before distributing it. The article describes this process at a high level: Project Zero privately reports vulnerabilities to vendors, giving them time to develop security updates and patches. It does not provide examples of code changes or describe testing methods.

Project Zero’s changing deadlines affect how much time vendors usually have before technical details are published. The team says AI has accelerated discovery and exploitation, while vendors warned that moving straight to 30 days could put users at risk. The gradual schedule gives vendors time to improve remediation systems, but the general rule is that details will be published after the deadline even if a patch is not ready.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web