Google Project Zero will cut its bug disclosure deadline from 90 days to 30 by 2028, citing AI-speed attacks

Science & Technology

Google Project Zero cuts vendor disclosure to 30 days by Jan. 3, 2028

Google Project Zero will shorten vendor disclosure windows from 90 to 60 days in 2027 and 30 days in 2028, citing AI-assisted exploit development and patch pressure.

By
Sophie Lindqvist
Filed
Channel
Science & Technology
Read
4 min read

By Sophie Lindqvist

Google Project Zero will reduce its vulnerability disclosure window from 90 days to 30 days by Jan. 3, 2028. The deadline falls to 60 days on Feb. 2, 2027, while the extra 30 days for patch rollout has already ended. Security engineer Natalie Silvanovich announced the change on Nov. 13.

What changes in Project Zero’s disclosure timetable?

Project Zero is Google’s in-house team for finding security flaws in other companies’ software. Under its “Full X” policy, it publishes vulnerability details X days after reporting a flaw to the vendor, whether or not a fix exists.

The schedule runs as “Full 90” from Nov. 13, “Full 60” from Feb. 2, 2027, and “Full 30” from Jan. 3, 2028.

Since 2021, the team has used a “90 + 30” policy. Vendors had up to 90 days to fix a reported flaw, followed by 30 days of “soak time” for users to install the patch. Project Zero will continue tracking fixes, but fix status will no longer decide when publication occurs.

That removes both the extension for unready vendors and negotiations over whether a defect qualified for additional time, two processes Google described as increasingly difficult to manage.

Why does Google consider AI the deciding factor?

AI-assisted discovery and exploitation has made the old timetable clearly insufficient, Silvanovich wrote. Large language models help attackers locate flaws, exploit them faster, and increase the chance that several parties, including malicious ones, discover the same defect.

Silvanovich wrote that reverse engineering now takes so little time that “zero patch-to-exploitation time is a reasonable, though imperfect, model of attacker behavior.” Google treats that model as removing the practical value of extra time for users to install a patch.

Google’s threat intelligence group described one such tool on Monday. The system converted security patches into working exploits without human involvement. The announcement did not identify the tool or publish an exploitation-time series.

“We are finding several vulnerabilities a week in targets where we once found a few vulnerabilities a year,” Silvanovich wrote. She said negotiations over fix status and grace periods were quickly becoming unmanageable.

The stronger discovery rate compresses the interval between vendor awareness and the prospect that attackers independently reach the same flaw, making fixed publication dates less reassuring than they once seemed.

Why did Google phase in the shorter deadlines?

Vendor feedback shaped the phase-in. Project Zero said an immediate cut could expose users to unfixed bugs, divert engineering teams toward new reports at the expense of severe older flaws, and discourage some vendors from reporting or fixing vulnerabilities at all.

Silvanovich said the policy encourages vendors to improve remediation time on a schedule that is challenging but realistic. The phase-in gives vendors time to shorten remediation cycles before the 30-day regime takes effect.

The team also rejected severity as the deadline-setting variable. Severity measures a flaw’s impact on users, not the probability that an attacker will find it, according to Project Zero.

That distinction keeps timing uniform but leaves vendors facing the operational burden of producing a tested fix, a cost the announcement did not quantify.

Which vulnerabilities escape the standard timetable?

Most special cases have disappeared, but several routes still bypass the normal clock:

  • Bugs already exploited in the wild: These follow Google’s application security disclosure policy, which SecurityWeek reports sets a seven-day window.
  • Publicly known flaws: Project Zero may publish early when details are already public.
  • Vendor requests: A vendor can ask the team to publish before the scheduled deadline.
  • Rare hardware cases: Project Zero will still grant exceptions for serious hardware flaws, as it did for Spectre and Meltdown.

What has Google not established?

The announcement gives no count of reported flaws in each exception category, no measured distribution of patch-to-exploit times, and no vendor-by-vendor readiness data. It supports the direction of the policy, not a precise forecast of exploit speed or disclosure volume.

Project Zero says its discovery pace rose from a few flaws a year in some targets to several a week, while Google treats near-instant exploitation as reasonable but imperfect.

By Jan. 3, 2028, vendors that remediate faster will carry less schedule risk, though Google has offered no comparative data on their readiness. The first test comes on Feb. 2, 2027, when the standard window shrinks to 60 days. Vendors must then show whether faster disclosure improves patch deployment without diverting attention from severe older flaws.

Source: mlq.ai

Share this article:

More from Sophie Lindqvist

Sophie Lindqvist

Show full bio

News editor covering money, funding and market shifts around Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,054 articles

Related articles

« Previous article