Google Drastically Shortens Vulnerability Fix Timelines in Response to AI Threats
—
# Google accelerates security vulnerability disclosure: goodbye to the grace period
In a major turning point for global cybersecurity, Google’s research team Project Zero has announced a radical reform of its vulnerability disclosure policy. The goal? To end what experts now call the “false sense of security” and force manufacturers to react with unprecedented speed.
Until now, the standard procedure offered publishers a comfortable window: 90 days to develop a patch, followed by an additional 30-day grace period before Google published the technical details of the flaw. This timeframe allowed users to install updates without excessive pressure.
Today, this obsolete model has been abandoned.
A new era imposed by artificial intelligence
The reason for this abrupt change lies in the rise of artificial intelligence. According to Natalie Silvanovich, an engineer at Project Zero, the advent of AI tools has upended the threat landscape.
“AI-assisted exploitation has transformed vulnerabilities,” she explains. Google researchers now discover several critical flaws per week in targets where they previously found only one or two per year. Meanwhile, attackers use AI to instantly analyze published patches to find other similar unpatched vulnerabilities.
In this context, the traditional 90-day deadline is deemed “clearly insufficient.” The American giant explains that even if a flaw is patched on time, publishing the patch allows hackers to decrypt the vulnerability and target those who have not yet updated their systems.
A tight schedule: 60 days, then only 30
Google details a two-phase action plan to accelerate the pace:
- **Starting February 2, 2027**: The fix deadline drops from 90 to **60 days**. The post-patch grace period is eliminated. Once the 60 days expire, technical details are made public, even if no patch is yet available.
- **Starting January 3, 2028**: The deadline is reduced again to just **30 days**. That is three times less time than today. Google commits to publishing information on the flaw as soon as this threshold is reached, regardless of the status of patches.
This gradual transition aims to avoid an immediate shock that could leave publishers defenseless while they reorganize their internal processes.
Industry reactions: between concern and adaptation
The decision does not have unanimous support within the cybersecurity community. Several experts fear the imposed pace is too fast for small development teams.
Katie Moussouris, CEO of Luta Security, emphasizes that “a deadline doesn’t magically produce a fix.” On her side, Dustin Childs from Trend Micro fears the emergence of rushed, poorly developed patches that could leave backdoors open.
To reassure the industry, Google cites its own statistics: all vulnerabilities reported by Project Zero in 2026 for open-source projects were fixed before the 30-day deadline (with a single exception repaired in 31 days). The giant claims this new cadence should not put manufacturers in difficulty.
What does this mean for you?
For the end-user, the consequences are twofold: faster and mandatory security updates, but also heightened vigilance. There is no longer a “shooting window” where a system appears secure while hackers already know about a flaw.
The lesson is clear: install updates as soon as they become available on your computer and smartphone; it is no longer an option but an absolute necessity. Do not rely on the existence of a patch to protect yourself; protect yourself before it is even published.
Other tech giants are following suit. Anthropic, creator of the Claude model, has also announced reducing its disclosure deadline for severe vulnerabilities to 30 days starting next January. The entire industry is thus preparing to enter an era where speed of correction will be the only guarantee of security.
—