16 Apr AI Cybersecurity raises stakes for global digital infrastructure 

A new generation of artificial intelligence models is transforming cybersecurity, with implications for governments, businesses, and critical infrastructure worldwide. 

In recent months, AI systems have evolved from producing unreliable outputs to identifying serious vulnerabilities in widely used software.  

These tools underpin essential technologies, making the shift particularly consequential for a region accelerating its digital transformation. 

The latest development comes from Anthropic, which this week revealed a powerful model called Mythos Preview.  

The company claims the system can detect high-severity security flaws across major operating systems and web browsers and even suggest methods to exploit them. 

While this marks a breakthrough for defensive cybersecurity, it also raises concerns about misuse by hackers or hostile state actors. 

For countries investing heavily in smart cities, digital banking, and energy infrastructure, the dual-use nature of such technology presents both opportunity and risk.  

Anthropic has restricted access to Mythos Preview, limiting it to around 50 organisations under a program known as Project Glasswing.  

The initiative aims to secure critical global software systems, including open-source technologies that underpin much of the internet. 

At the same time, AI is already improving defensive capabilities. However, its rapid improvement has not been without challenges.  

Just a year ago, developers were overwhelmed by inaccurate, AI-generated bug reports. Now, the situation has reversed, as reports are more precise and increasingly valuable, accelerating the pace of security fixes. 

Despite this progress, experts warn that AI remains better at identifying problems than solving them.  

Fixing vulnerabilities still requires human judgment, especially in complex systems where security decisions involve trade-offs. 

The broader concern lies in the global race to develop and deploy advanced AI models.  

For policymakers, this raises urgent questions about regulation, collaboration, and cyber defence readiness.  

As AI capabilities continue to advance, the balance between technological innovation and security will become a defining issue for the region’s digital future.