Beyond the AI Doom Loop: A New Era of Cybersecurity
As part of our engineering podcast series, hosted by David Huish, Head of Engineering at Aspiron Search, we speak with engineering leaders and experts working at the forefront of cybersecurity to explore the technologies, challenges and ideas shaping the industry.
David was recently joined by Matt Suiche from Tolmo to discuss how AI is changing the cybersecurity landscape, from making vulnerability discovery faster to helping uncover years of hidden technical debt.
With almost 20 years of experience across software engineering, security and incident response, Matt shares why the rise of AI doesn't have to be all doom and gloom, and how it could help create a new era of cybersecurity.
AI is changing the cybersecurity landscape
AI has developed rapidly over the last couple of years. What started with basic coding assistance has quickly evolved into agents capable of generating entire programmes, researching documentation and carrying out complex tasks.
For Matt, the real turning point came when AI started becoming more widely used for coding and software development.
As more people use coding agents to build and automate things, the number of potential vulnerabilities and security risks is also increasing.
Attackers are already benefiting from better automation, allowing them to discover vulnerabilities and build more complex attack chains much faster than before.
The challenge for defenders is keeping up with that pace.
Why businesses need to rethink their approach to security
One of Matt's biggest concerns is that businesses can no longer assume their systems are completely secure.
With vulnerabilities appearing across infrastructure and widely used open-source software, simply identifying and patching critical issues isn't enough.
Instead, organisations need to think about what happens when an attacker does get in.
Matt highlighted the importance of designing systems with the assumption that a breach could happen. This means looking at areas such as network segmentation, reducing attack surfaces and making sure vulnerabilities can be contained before they spread across the wider business.
It's about moving beyond simply preventing attacks and thinking about how to limit the damage when something goes wrong.
The positive side of AI: uncovering years of technical debt
While much of the conversation around AI and cybersecurity focuses on the risks, Matt believes there's a significant positive side to what's happening.
AI is giving security researchers more opportunities to find vulnerabilities that have been sitting in software for years, sometimes without anyone noticing.
Previously, finding these issues required a huge amount of manual effort and specialist knowledge. Now, AI can help more people investigate code, identify weaknesses and uncover problems that may have gone unnoticed.
Matt describes this as the gradual removal of decades of technical debt.
He also believes we're likely to see more software being rewritten with security in mind, including greater adoption of memory-safe programming languages such as Rust.
The result? More eyes on existing systems, more vulnerabilities being identified and, hopefully, more secure infrastructure in the long run.
Everyone is becoming a builder
Another interesting shift Matt highlighted is how AI is changing who builds software.
Previously, developing tools and applications was largely the responsibility of software engineers. Now, coding agents are giving people across different departments the ability to build their own solutions.
This includes people in marketing, operations and security who might previously have relied entirely on off-the-shelf software.
But this also creates new challenges.
As more people build their own tools and applications, security becomes everyone's responsibility, rather than something that sits separately with a dedicated team.
Matt sees a shift towards developers taking greater ownership of the security of what they build, with the traditional boundaries between software engineering and cybersecurity becoming increasingly blurred.
What does this mean for the future of cybersecurity?
Despite the headlines, Matt doesn't believe businesses should panic about the pace of AI development.
For security leaders, much of the approach remains the same: assume a breach could happen, focus on segmentation, keep systems patched and make sure the right controls are in place.
He also believes AI will become essential for defenders. With the volume of vulnerabilities and security events increasing, it's becoming increasingly difficult to manage everything manually.
Over the next six to twelve months, Matt expects to see the gap between attackers and defenders begin to close as security teams increasingly adopt AI themselves.
For anyone looking to enter the cybersecurity industry, his advice is to focus on developing a strong problem-solving mindset. Technical knowledge is still valuable, but being able to approach problems differently, think creatively and spot opportunities will become increasingly important.
The agentic era is bringing new challenges, but it's also creating opportunities to build more secure systems. The businesses that embrace these changes while taking security seriously will be in a much stronger position as the industry evolves.
Check out the full podcast here!
Frequently Asked Questions
1. How is AI changing cybersecurity?
AI is making it easier to discover vulnerabilities, automate attacks and build more complex attack chains. It's also helping security teams identify weaknesses, analyse systems and improve their ability to detect and respond to threats.
2. What is agentic AI security?
Agentic AI security focuses on protecting AI agents, the systems they interact with and the actions they can take. This includes managing permissions, controlling access and preventing agents from creating new security risks.
3. Why is technical debt a cybersecurity risk?
Technical debt refers to outdated systems, software and code that may contain weaknesses or vulnerabilities. Over time, these can create security risks if they aren't properly maintained or updated.
4. Why is network segmentation important in cybersecurity?
Network segmentation separates systems and resources into smaller sections. If an attacker gains access to one area, segmentation can help prevent them from moving freely across the wider network.
5. Will AI make cybersecurity more important?
Yes. As more businesses adopt AI and use coding agents to build applications, the potential attack surface grows. Security will increasingly need to be considered throughout the development process, rather than treated as a separate task.
6. What skills will be important for the next generation of cybersecurity professionals?
Problem-solving, critical thinking and the ability to approach challenges creatively will become increasingly important. As AI takes on more technical tasks, understanding how to identify problems and find new ways to solve them will be valuable.