Walking the Tightrope of AI Implementation

In today’s rapidly evolving tech landscape, the decision to implement AI presents a classic dilemma for businesses. On one hand, delaying adoption might mean missing out on significant productivity gains and competitive advantages. On the other, a hasty or improper implementation could expose businesses to substantial risks. Enter a new breed of startups focused on “security for AI,” tackling these very threats head-on.

Meet the Trailblazers in AI Security

Among these pioneering firms is Mindgard, a British university spin-off joining ranks with Israeli startup Noma and U.S.-based Hidden Layer and Protect AI. As Professor Peter Garraghan, CEO and CTO of Mindgard, articulates, “AI is still software, so all the cyber risks you probably heard about also apply to AI.” However, he points out that the opaque and inherently unpredictable nature of neural networks necessitates a novel approach.

The Innovative Approach of Mindgard

Mindgard’s strategy leverages Dynamic Application Security Testing for AI (DAST-AI), focusing on vulnerabilities that only emerge during runtime. This involves continuous and automated red teaming—simulating attacks using an extensive threat library. For example, it tests the resilience of image classifiers against adversarial inputs. This cutting-edge technology draws from Garraghan’s extensive background in AI security research.

“We built this company to do positive good for the world, and the positive good here is people can trust and use AI safely and securely.”

Professor Peter Garraghan

Building Bridges: From Academia to Industry

The dynamic nature of AI security means that staying ahead requires constant innovation. Garraghan believes that Mindgard’s strong ties with Lancaster University provide a unique advantage. The firm will own the intellectual property generated by 18 doctoral researchers over the coming years—a deal unmatched globally.

Scaling Up with Strategic Investments

Mindgard is making significant strides in commercializing its offering as a SaaS platform. With co-founder Steve Street driving operations, the company targets enterprises and red teams needing robust AI risk prevention solutions. To support its expansion into the U.S., Mindgard recently raised $8 million led by Boston-based .406 Ventures.

• Funding will bolster team growth, product development, and U.S. market expansion.
• The company’s R&D will remain in London, maintaining its innovative edge.
• With plans to grow modestly from 15 to 20-25 staff by year-end, Mindgard remains agile.

A Future Ready for AI Security

Professor Garraghan envisions a world where AI is deployed ubiquitously—and securely. While AI security is still maturing, companies like Mindgard are laying critical groundwork to ensure that when security threats arise alongside widespread AI adoption, they are ready to tackle them head-on.