Google and Anthropic are slashing their coordinated vulnerability-disclosure timelines, compressing traditional 90-day embargoes down to as little as 30 days in a sweeping industry response to AI-accelerated vulnerability discovery and exploitation.
The policy overhauls, announced independently this week by Google Project Zero and Anthropic, abandon the assumption that software flaws stay safely hidden for months while vendors develop and deploy patches.
With frontier models capable of autonomously discovering, reverse-engineering, and weaponizing software flaws in hours, both labs concluded that long disclosure delays now give attackers a predictable blueprint while providing defenders a false sense of security.
Google’s current “90+30 days” policy gives vendors 90 days to remediate a vulnerability discovered by its researchers, followed by extra 30 days of patch “soak time.” Beginning in February 2027, Google will drop the remediation window to 60 days; it will tighten that to 30 days by January 2028.
The overhaul scraps grace periods and fix-verification negotiations, as the surging volume of vulnerabilities makes manual tracking unmanageable.
Meanwhile, starting Jan. 1, 2027, Anthropic is shortening its disclosure window for vulnerabilities discovered by its frontier models to 30 days for high- and critical-severity issues and 45 days for all other vulnerabilities. It will reduce its post-patch technical publication delay from 45 days to 30 days.
However, Anthropic said it would retain flexibility for complex environments, offering tailored timelines for critical infrastructure, while maintaining a seven-day window for actively exploited flaws.
Both labs pointed to AI as the primary catalyst. Natalie Silvanovich, security engineer at Google Project Zero, noted that LLMs have fundamentally altered the threat landscape.
“A vulnerability reported today presents a much higher immediate risk to users than one reported last year,” she said. “Aspects of our policy that require vendor negotiation or agreement, including determining whether a bug is fixed and whether a grace period applies, are quickly becoming unmanageable.”
Anthropic’s Frontier Red Team echoed the sentiment regarding independent rediscovery, saying in a blog post “A bug that one model can find is increasingly likely to be found by another. We are seeing a higher rate of vulnerability rediscovery as these capabilities increase.”
Vendor pressure and institutional reality
While the moves are aimed at encouraging faster remediation across the software ecosystem, they “place greater pressure on vendors to address vulnerabilities,” according to Ramy Rahman, senior principal solutions engineer at vulnerability-management firm ArmorCode.
“To respond to these new timelines, organizations should be better prepared to assess whether compensating controls can reduce the risk of a vulnerability when it’s disclosed. With disclosures arriving sooner, and sometimes before a fix is available, teams need a process for evaluating exposure, applying available mitigations, and tracking remediation.”
Neil Osnato, founder of Persistence Analytics Group, said accelerating the disclosure clock exposes a dangerous institutional fallacy. “Disclosure is an information event, but remediation is a technical event,” he told TechTarget, noting that operational deployment is an entirely separate institutional decision.
“Faster disclosure does not necessarily mean faster risk resolution,” Osnato said. “An affected institution may still be operating equipment or software that has not been successfully remediated, because of compatibility testing, third-party dependencies, maintenance windows, legacy infrastructure, or the consequences of taking a critical system offline.”
The automated threat horizon
The news comes as Google warned this week that AI threats have crossed a “significant threshold,” with attackers now fully automating vulnerability discovery and exploitation.
A new Google Threat Intelligence Group report said threat actors are using LLM frameworks to fully automate vulnerability research and exploitation, with no human in the loop.
The report reinforces growing concerns about the accelerated exploitation of n-day vulnerabilities – known flaws that often remain unpatched for months or years in enterprise environments.
“Traditional patching cycles are becoming obsolete, and some fixes will need to happen in hours rather than days or weeks,” said Jason Elrod, CISO at MultiCare Health System.
As defenders race to compress their remediation timelines, AI labs argue that shorter disclosure windows are ultimately meant to drive industry-wide resilience. Michael Moore, cybersecurity lead at Anthropic, said proactive disclosure is designed to help defenders stay ahead of machine-speed threats.
“Frontier models have become much better at finding vulnerabilities,” Moore said in a LinkedIn post. “I’m optimistic that this helps defenders fix them faster than ever. But what one model finds, another increasingly will too.”
James Walker is lead editor at TechTarget Cybersecurity.