The cybersecurity research team will further reduce the disclosure period to 30 days by January 2028, citing faster AI-driven attacks, patch reverse-engineering and a growing volume of vulnerabilities.
Google’s Project Zero, the company’s internal cybersecurity research team, plans to progressively shorten the time vendors have to address reported security vulnerabilities before public disclosure. Under its revised policy, the disclosure window will fall from the current 90 days to 60 days from February 2, 2027, and then to 30 days from January 3, 2028.
The policy change reflects growing concerns that attackers are increasingly capable of discovering and exploiting software vulnerabilities at speed, partly driven by advances in large language models (LLMs) and artificial intelligence.
Project Zero said vulnerabilities reported today pose a greater immediate risk to users than those reported in previous years, as multiple parties, including malicious actors, may independently identify and exploit the same security flaw.
New Vulnerability Disclosure Timelines
Since 2021, Project Zero has followed a “90 + 30” disclosure policy. It allowed vendors up to 90 days to fix a reported vulnerability, followed by an additional 30-day period, known as “soak time”, for the patch to reach users.
The revised policy replaces this approach with a single disclosure window, with the following schedule:
- Current policy: Vulnerabilities are disclosed 90 days after being reported to the vendor.
- February 2, 2027: The disclosure window will be reduced to 60 days.
- January 3, 2028: The disclosure window will be reduced further to 30 days.
The revised timeline is intended to encourage faster remediation and reduce the period during which users remain exposed to known but unpatched vulnerabilities.
Project Zero acknowledged that shorter deadlines could create challenges for vendors. An overly aggressive timeline may cause companies to prioritise newly reported vulnerabilities over more severe issues, while deadlines they cannot realistically meet could discourage remediation efforts.
The team said it had selected a transition schedule that it considers “challenging but realistic”.
Why Google Is Shortening the Disclosure Window
Project Zero cited three key developments behind the policy revision.
1. AI is accelerating vulnerability discovery and exploitation
The growing capabilities of LLMs have increased the speed at which attackers can identify and exploit software vulnerabilities, according to Project Zero.
The technology also increases the possibility that different parties may discover the same vulnerability independently, raising the likelihood that a flaw could be exploited before a vendor has time to address it.
2. Attackers can reverse-engineer security patches
Project Zero said attackers can reverse-engineer software patches almost immediately after their release to identify the underlying vulnerability and develop exploits.
As a result, the team believes the traditional soak-time period is no longer a reliable security measure. It described the additional window as “no longer meaningful” and potentially creating “a false sense of secrecy”.
The focus, therefore, is shifting towards reducing the “patch gap” — the time between a security fix being released and users installing it. Vendors will need to accelerate patch distribution or introduce additional protections while a permanent fix is pending.
3. Rising vulnerability volumes are making negotiations difficult
Project Zero also pointed to an increase in the number of vulnerabilities it identifies. The team said it now discovers several vulnerabilities in a week in some targets where it previously found only a few in a year.
Managing individual negotiations with vendors over remediation status, disclosure dates and additional grace periods is becoming increasingly difficult, it added.
Project Zero Changes Its Approach to Vendor Remediation
The revised policy will also change how Project Zero evaluates whether a vulnerability has been fixed.
The team will no longer use the question of whether a bug has been fixed as a key factor in determining its disclosure timeline. It will continue tracking remediation efforts and identifying vulnerabilities that remain unresolved, but will avoid making a binary judgement that carries consequences for vendors.
This means vendors will no longer be able to delay a disclosure simply by reporting that a fix is in place. The treatment of partial mitigations under the revised approach, however, remains unclear.
Project Zero said it sought feedback from vendors that frequently receive its reports. Some warned that an immediate reduction in disclosure timelines could expose users to unfixed vulnerabilities while companies adapt their security and patch-management processes.
Google has consequently opted for a phased implementation. The policy also assumes that Project Zero’s reports are reproducible, readable and accessible, while recognising that vendors may respond differently to reports that do not meet these standards.
Google Flags Rise of Autonomous AI-Driven Cyberattacks
The policy revision comes alongside a separate report from Google Threat Intelligence Group (GTIG), which identified a shift towards more autonomous, AI-enabled cyber operations.
In its AI Threat Tracker report, GTIG said threat actors are moving beyond manually prompting AI systems towards autonomous, multi-agent frameworks capable of carrying out multiple stages of cyberattacks.
For the first time, the group identified threat actors using LLM-based frameworks to discover and exploit vulnerabilities without a human actively directing the process.
According to GTIG’s data, disclosures and exploits nearly doubled between January and August 2026. The report also recorded an increase in average monthly zero-day exploits, from eight in 2025 to 11 in 2026, before the figure rose to 22 in August 2026.
The findings point to a changing cybersecurity environment in which AI tools can help attackers scale operations and act more quickly against known security weaknesses.
Google Calls for Greater Focus on Agentic Cyber Defence
John Hultquist, Chief Analyst at Google Threat Intelligence Group, said attackers are not necessarily relying on entirely new techniques to compromise systems. Instead, they are increasingly exploiting known vulnerabilities that remain unpatched, but doing so with greater speed and scale.
He argued that organisations will need to embrace “agentic defense” to keep pace with this shift in the threat landscape.
The revised Project Zero policy reflects a broader move towards faster vulnerability remediation, shorter exposure periods and more responsive security practices. As AI tools make vulnerability discovery and exploitation more accessible, the ability of vendors to distribute patches quickly and protect users during the remediation process is likely to become increasingly important.
By : Vanshika Tayal




