Is AI Security the Next Frontier in Tech? How one startup is tackling the hidden risks of artificial intelligence

As AI systems become more powerful, so do the threats they pose - from data breaches to manipulation. Peter Garraghan, co-founder of Mindgard, a London based 'security for AI' startup is on a mission to safeguard organisations from these evolving risks, blending cutting-edge research with real-world solutions to ensure AI adoption remains both secure and ethical.

What inspired you to start your business?
Mindgard was born from a desire to bridge academic research and real-world impact. As a professor specializing in computing systems, AI security, and machine learning, I saw firsthand the immense risks AI systems pose, from data leaks to manipulation. I knew traditional application security tools would struggle to address these vulnerabilities, so I founded Mindgard to create innovative solutions that protect organizations and enable safe AI adoption.

What was your biggest challenge and how did you overcome it?
We operate within an emerging market segment where researchers and engineers are only scratching the surface of risks within AI systems. This has meant a lot of education to both customers and investors, although the conversations have shifted away from hypotheticals towards more grounded problems, as evidenced by the flaws in Microsoft's AI filters that we discovered last year.

How did you secure your initial funding?
We officially founded Mindgard in May 2022, with invaluable support from Lancaster University. However, creating a university spin-out required more than research skills—it meant raising capital, refining our value proposition, and preparing the technology for demos, all while balancing my role as a professor. Spin-outs thrive not just on groundbreaking technology but on addressing real business needs and delivering value that attracts and retains customers. This approach helped us secure critical funding to drive Mindgard forward.

Related: The Journey to Centaur Status

How do you handle failure or setbacks?
Setbacks are part of working in a field as dynamic as AI security. The rapid pace of change in frontier models and AI development architectures often forces us to reevaluate and adapt our approach. While the fast-moving technology and talent gaps present a challenge, they also drive us to stay ahead of the curve and continually innovate. We handle these challenges by remaining agile, focusing on education, and iterating quickly on our strategies. Every setback is an opportunity to reinforce the importance of proactive AI security solutions.

What advice would you give to someone starting their own business?
Focus on solving real problems, and better yet quantify the size of the problem. It's not enough to have a groundbreaking idea; it must deliver tangible value, and must be appropriately scoped to go after. Become versed in talking about the negative consequences of potential customers sticking with the status quo, and likewise the positive outcomes you will deliver to your customers.

How do you stay motivated during tough times?
Motivation comes from the mission - securing the world's AI. Knowing that our work secures organizations and enables safer AI adoption keeps me driven. Seeing the tangible impact of Mindgard's solutions, whether through partnerships or safeguarding our customers' critical systems, reinforces the importance of what we do.

Share your tips for achieving success.....
Success comes from persistence, adaptability, and a clear vision. Build a team of diverse talents who challenge your ideas and bring complementary skills to the table. Focus on solving real problems and delivering value while being open to learning and growth. Each challenge you face is a stepping stone toward achieving your goals, so celebrate progress and use it to fuel momentum for the next milestone.

Related: Rethinking Carbon Credits: A New Era for Climate Action