Mindgard, a London-based DeepTech startup specialising in security for AI, has raised £3 million in seed funding.

Initially developed under the aegis of Lancaster University, Mindgard harnesses years of PhD-level research in AI security with their automated threat analysis, detection, and response platform already in use by the intelligence community. 

The Mindgard platform fully integrates with market-leading cybersecurity products, and is a part of the NVIDIA Inception Programme and Microsoft Founder’s Hub.

The funding comes from IQ Capital and Lakestar and will be used to bring the platform to market. 

“Today’s businesses are rapidly embracing LLMs (large language models) and other AI technologies. Although these enterprises understand the need to secure their AI assets, the cybersecurity that worked for them in the past won’t help guard their AI against the next generation of cyber threats,” said Dr Peter Garraghan, CEO at Mindgard and Professor at Lancaster University.

“This leaves companies open to the possibility of AI models and data being stolen or compromised, as well as disabling attacks from bad actors.

“Currently, Mindgard is deployed by key intelligence organisations around the world to assess, detect, and respond to AI issues and attacks as they happen. With this seed funding, Mindgard now becomes available to enterprises who need to protect their operations against the rapidly-evolving AI threat landscape. 

“We are defining and driving the security for AI space, and believe that Mindgard will quickly become a must-have for any enterprise with AI assets.” 

Exploitation is AI’s elephant in the room

Archie Muirhead, partner at IQ Capital, added: “AI models are an emerging critical attack surface for organizations increasingly seeking to deploy AI.

“Mindgard provides the most comprehensive solution to manage this surface and reduce risk. With Mindgard’s platform, the complexity of model assessment is made easy and actionable through integrations into common MLOps and Sec Ops tools and an ever-growing attack library. 

“We are happy to back a remarkable founding team spinning out of a world-leading AI security lab at Lancaster.”

The role of diversity and inclusion in tackling biased AI