The increasing adoption of artificial intelligence (AI) in various industries has created a dilemma for companies: implement AI and risk exposure to unmitigated threats, or wait and potentially miss out on productivity gains and benefits. This is where a new wave of "security for AI" startups comes in, aiming to mitigate the risks associated with AI deployment. One such startup is British university spinoff Mindgard, which offers a unique approach to AI security.

Mindgard's CEO and CTO, Professor Peter Garraghan, emphasized that AI is still software, and therefore, all traditional cyber risks apply. However, the opaque nature and intrinsically random behavior of neural networks and systems justify a new approach. This is where Mindgard's Dynamic Application Security Testing for AI (DAST-AI) solution comes in, targeting vulnerabilities that can only be detected during runtime.

The DAST-AI solution involves continuous and automated red teaming, simulating attacks based on Mindgard's threat library. For instance, it can test the robustness of image classifiers against adversarial inputs. Garraghan's background as a professor and researcher focused on AI security has been instrumental in developing this technology, which is fast evolving to keep pace with the rapidly changing landscape of large language models (LLMs) and threats.

Mindgard's ties to Lancaster University will also play a crucial role in keeping the company at the forefront of AI security research. The startup will automatically own the IP to the work of 18 additional doctorate researchers for the next few years, a unique arrangement that Garraghan believes is unparalleled in the industry.

Despite its research roots, Mindgard is a commercial product, offering a SaaS platform led by co-founder Steve Street as COO and CRO. The company is already working with enterprises, traditional red teamers, and pen testers, as well as AI startups that need to demonstrate AI risk prevention to their customers.

To further accelerate its growth, Mindgard has announced an $8 million funding round led by Boston-based .406 Ventures, with participation from Atlantic Bridge, WillowTree Investments, and existing investors IQ Capital and Lakestar. This follows a £3 million seed round in 2023. The new funding will be used to build the team, develop the product, and expand into the U.S. market.

As part of its expansion plans, Mindgard has appointed a VP of marketing, Fergal Glynn, who is based in Boston. However, the company plans to maintain its R&D and engineering operations in London. With a current headcount of 15, Mindgard aims to reach 20 to 25 people by the end of next year, remaining relatively small as it prepares for the anticipated growth of the AI security market.

According to Garraghan, Mindgard's mission is to enable people to trust and use AI safely and securely. As AI deployment becomes more widespread, the need for robust security solutions will only increase. With its innovative approach and strong research backing, Mindgard is well-positioned to capitalize on this trend and make a positive impact on the world.