Artificial intelligence is becoming one of the most important tools in modern cybersecurity and Anthropic is taking another major step toward putting advanced AI capabilities directly into the hands of cybersecurity defenders.
Anthropic has announced an expanded Cyber Verification Program (CVP) designed to give qualified cybersecurity professionals greater access to its most capable Claude models for legitimate defensive security work.
The expanded program introduces three access tiers, allowing security teams to receive access appropriate to the type and sensitivity of cybersecurity work they perform. The program includes advanced models such as Claude Opus 5.5, Claude Sonnet 5.5 and Claude Mythos 5.1, along with future models.
Why Anthropic Is Expanding the Program
Cybersecurity is entering a new era where both defenders and attackers can use AI to automate complex tasks.
AI can help security teams analyze source code, identify vulnerabilities, investigate suspicious activity, test applications and develop patches. At the same time, increasingly capable AI systems could potentially help attackers discover vulnerabilities and develop attacks much faster.
Anthropic says its earlier Project Glasswing initiative demonstrated the potential of advanced AI for defensive cybersecurity.
According to Reuters, organizations participating in Project Glasswing identified at least 129,000 verified software vulnerabilities between April and July, while Anthropic’s own open-source scanning identified thousands more. More than 33,000 of the reported vulnerabilities were classified as critical or high severity.
The numbers demonstrate why AI-assisted vulnerability discovery could become an important part of future cybersecurity operations.
From Project Glasswing to a Broader Cybersecurity Program
Project Glasswing was originally created to provide selected organizations with access to Anthropic’s highly capable cybersecurity AI. The initiative brought together major technology and security organizations to identify and fix vulnerabilities in critical software and infrastructure. Anthropic has now incorporated Glasswing into the expanded Cyber Verification Program.
Existing Glasswing participants will transition to the program’s Specialized Access tier. The new structure is designed to make advanced defensive capabilities available to a broader group of legitimate security professionals while maintaining controls around highly sensitive cyber capabilities.
Why This Matters for Cybersecurity
Traditional vulnerability assessment can require significant amounts of time and highly specialized security expertise but AI is changing that equation.
An advanced AI system can analyze large amounts of code, identify unusual patterns, reason through potential attack paths and help security teams prioritize vulnerabilities.
For cybersecurity teams, this could mean:
- Faster vulnerability discovery
- Automated code security analysis
- More efficient penetration testing
- Faster security investigations
- Improved vulnerability prioritization
- Faster development of security patches
- Greater protection for open-source software
Anthropic previously said that its goal was to use increasingly powerful AI capabilities to help defenders find and fix vulnerabilities before comparable capabilities become widely available to malicious actors.
AI Could Change the Role of the Cybersecurity Professional
The growth of AI-powered cybersecurity does not necessarily mean cybersecurity professionals will become less important. Instead, their roles are likely to change.
Security analysts may increasingly use AI to investigate alerts and summarize large amounts of security data. Penetration testers could use AI to identify potential attack paths more efficiently. Security engineers could use AI to review code and recommend fixes.
This means future cybersecurity professionals will need to understand both security fundamentals and AI-assisted security tools.
Skills such as SIEM + AI | EDR + AI | Vulnerability Management + AI | Threat Intelligence + AI | Cloud Security + AI | Identity Security + AI could become increasingly valuable.
The Security Challenge
There is also an important risk. The same AI capabilities that help defenders find vulnerabilities can potentially help attackers discover weaknesses faster.
This creates an accelerating cybersecurity race. If an AI system can discover a vulnerability in hours instead of weeks, organizations have less time to patch their systems. Security teams therefore need better vulnerability management, faster patching processes, stronger identity controls and continuous monitoring.
Anthropic’s expansion reflects this reality: organizations cannot simply wait for AI-powered cyber threats to become widespread before preparing for them.
What This Means for Businesses
Businesses should begin treating AI capability as part of their cybersecurity strategy. Organizations should consider:
1. AI-assisted security operations
Use AI to help analysts investigate alerts, summarize incidents and identify suspicious patterns.
2. Automated vulnerability management
Use AI-assisted tools to continuously examine applications and infrastructure for weaknesses.
3. Secure software development
Integrate AI-powered security testing into the software development lifecycle.
4. Strong identity security
Strengthen MFA, Conditional Access, privileged access management and identity monitoring.
5. AI governance
Establish policies governing how employees and security teams can use AI with sensitive company information.
The Bigger Picture
Anthropic’s latest announcement is more than another AI product update. It signals a broader shift in cybersecurity. AI is becoming a security capability that organizations will increasingly need to deploy, manage and govern.
The question is no longer whether AI will affect cybersecurity. It already does. The bigger question is whether defenders can adopt AI quickly enough to maintain an advantage over increasingly sophisticated attackers.
Anthropic’s expanded Cyber Verification Program is one attempt to address that challenge by putting advanced AI capabilities into the hands of vetted cybersecurity professionals.
For cybersecurity students, SOC analysts, penetration testers, cloud security engineers and security architects, the message is clear: AI is becoming another essential layer of the modern cybersecurity skill set.