New Zealand joins Anthropic’s expanded Project Glasswing as AI-driven vulnerability discovery accelerates cyber risk.
New Zealand is set to gain access to Anthropic’s expanded Project Glasswing program as governments respond to the growing cybersecurity impact of advanced AI models.
Claude Mythos Preview, Anthropic’s cybersecurity-focused AI model, has already helped partners identify more than 10,000 high- or critical-severity software flaws. Cloudflare reported thousands of bugs, while Mozilla fixed hundreds of Firefox vulnerabilities found during testing.
Project Glasswing initially included major technology and infrastructure companies, but Anthropic is now extending access to about 150 additional organizations across more than 15 countries, including New Zealand.
The expansion comes as cybersecurity teams face a growing gap between how quickly AI can find vulnerabilities and how long it takes humans to patch them. Anthropic has warned that similar models could become more widely available within months, potentially giving attackers access to faster and cheaper tools.
New Zealand is already strengthening its cyber resilience plans. Its Cyber Security Strategy 2026–2030 includes measures focused on critical infrastructure, risk management and stronger coordination between government and industry.
For local businesses, the message is clear: annual reviews and slow patch cycles may no longer be enough. As AI accelerates vulnerability discovery, organizations will need faster detection, tighter software inventories and more consistent security controls.
