TrendAI has joined Anthropic's Project Glasswing, a programme available to selected organisations in Australia and New Zealand.
It will use Anthropic's Claude Mythos Preview for vulnerability detection and remediation, with a focus on reviewing and analysing software code. The aim is to help threat intelligence researchers move faster from vulnerability discovery to disclosure and remediation.
Project Glasswing is designed to help organisations identify and address weaknesses in critical software systems. TrendAI said the work will also support vulnerability shielding and virtual patching as organisations seek to reduce risk across software environments.
Security focus
The move places TrendAI among a wider group of organisations participating in Anthropic's initiative. Participants are using advanced AI models to examine how they can support defensive cybersecurity work and improve the security of critical software infrastructure.
For TrendAI, the project extends its work in threat intelligence and defensive security as AI tools reshape both attack methods and protection efforts. It said AI is accelerating the discovery of software vulnerabilities, changing the pace at which defenders must respond.
That shift has become a central issue for businesses and public sector organisations that rely on large, complex software systems. Faster discovery of flaws can help defenders when tied to coordinated disclosure and patching, but the same advances also raise concerns that malicious actors could use similar tools to find and exploit weaknesses more quickly.
TrendAI said its work in the programme will focus on turning faster vulnerability discovery into practical defensive action. That includes helping organisations decide which issues to prioritise and how to reduce exposure before permanent fixes are applied.
Defensive AI
Mick McCluney, ANZ Field CTO at TrendAI, described the broader implications of frontier AI for the cybersecurity sector.
"Frontier AI models represent the most significant shift cybersecurity has ever seen, capable of discovering and exploiting vulnerabilities autonomously, at a scale and speed beyond human capability. The stakes couldn't be higher. TrendAI was built for exactly this moment. Our collaboration with Anthropic on Project Glasswing brings together TrendAI's 30 years of proven security expertise with the most advanced AI models in the world to strengthen Claude Mythos Preview and set a new standard for how frontier AI can be used in defence of our most critical sectors across Australia, New Zealand, and the Pacific Islands," said Mick McCluney, ANZ Field CTO, TrendAI.
Senior leadership at the company also framed the partnership as part of a broader effort to improve software security earlier in the development and deployment cycle. That approach reflects a growing industry focus on finding vulnerabilities sooner, before they can affect systems used in finance, government, healthcare, and other sectors that depend on critical digital infrastructure.
Earlier detection
"We're aligned with Anthropic's goals of using AI to make all software more secure. Organisations increasingly depend on software that operates at tremendous scale and supports critical business functions. Project Glasswing represents an important opportunity to explore how advanced AI can help software providers identify vulnerabilities earlier and improve the security and resilience of the systems customers depend on every day," said Rachel Jin, Chief Platform and Business Officer, Head of TrendAI.
Anthropic describes itself as an AI safety and research company, and its Claude family of models is used in areas including code understanding and security analysis. TrendAI's inclusion in Project Glasswing adds a cybersecurity specialist with a long-standing presence in threat research to that effort.
TrendAI operates as the enterprise AI security business unit of Trend Micro and says it works with organisations in 185 countries. In Project Glasswing, its role will be judged by how effectively advanced AI can help security teams find software flaws, disclose them responsibly, and reduce the window of exposure before attackers can take advantage of them.
TrendAI said insights from the programme will feed into broader industry efforts to strengthen the security of the digital ecosystem.