Everyone’s talking about AI making cyberattacks worse, painting a picture of rogue algorithms breaking down digital doors. But what if the real story is a bit more nuanced? What if the very technology we fear is also our best defense?
It’s easy to get caught up in the fear. The idea of AI-powered cyberattacks sounds like something out of a sci-fi movie, with super-intelligent programs finding every weakness in our digital infrastructure. The truth is, AI does present new challenges to cybersecurity. Robert Kim, MBA, recently summed up the impact of AI on cybersecurity in one word after the 2026 RSAC Conference: “impact.” That word choice might seem simple, but it speaks volumes about the significant shift happening.
The digital world runs on software. From the apps on our phones to the systems that control power grids, software is everywhere. If that software isn’t secure, bad actors can cause serious problems. Now, with AI becoming more common, the ways these bad actors might try to exploit weaknesses are changing. AI can potentially accelerate the discovery of vulnerabilities or create more convincing phishing attempts.
Project Glasswing Takes Flight
Recognizing this evolving threat, Anthropic launched Project Glasswing in 2026. This initiative isn’t just about playing defense; it’s about building a new kind of shield, one specifically designed for the AI era. The goal is clear: secure the world’s most critical software against AI-powered cyberattacks.
Project Glasswing isn’t a solo effort. Anthropic brought together some major players in the tech world. The list of collaborators includes:
- Amazon Web Services
- Anthropic
- Apple
- Broadcom
- Cisco
- CrowdStrike
This collaboration shows that securing our digital future against new threats requires a united front. These companies are working together to put in place cybersecurity measures tailored for AI’s unique challenges and opportunities.
NIST Steps Up
Government bodies are also getting involved. In 2026, the National Institute of Standards and Technology (NIST) released its preliminary draft of the Cyber AI Profile. This guidance helps map specific cybersecurity considerations related to AI. It’s a crucial step in creating a common understanding and framework for how we approach AI cybersecurity. This profile helps organizations identify and manage risks associated with using AI systems, as well as how to defend against AI-assisted attacks.
Think of it this way: just as we developed firewalls and antivirus software for traditional cyber threats, we now need new tools and strategies for AI. NIST’s guidance provides a starting point for developing these new approaches, helping organizations build more solid defenses.
AI as the Solution, Not Just the Problem
Here’s where my opening claim comes in. While AI certainly introduces new attack vectors, it also offers powerful tools for defense. AI can analyze vast amounts of data to detect unusual patterns that might signal an attack far faster than any human. It can automate responses to common threats, freeing up human experts for more complex issues. AI can even help identify vulnerabilities in software before attackers find them.
Project Glasswing, with its focus on securing critical software, acknowledges this dual nature of AI. It’s not just about protecting against AI threats, but also about using AI to build stronger protections. By understanding how AI can be used for both offense and defense, we can move beyond simply fearing the technology and instead work to use its capabilities for good.
The future of cybersecurity will likely involve a continuous back-and-forth between AI-powered attacks and AI-powered defenses. Initiatives like Project Glasswing are essential because they bring together the brightest minds to prepare for this future, making our digital world safer for everyone.
🕒 Published: