Safe Intelligence, a London-based leader in AI safety, has secured €5 million in a seed funding round aimed at transforming the reliability and security of artificial intelligence systems. The round, led by Amadeus Capital Partners, also saw contributions from OTB Ventures and existing investor Vsquared Ventures. With this funding, Safe Intelligence aims to bridge the gap between rapid AI development and the critical need for reliability in industries such as finance, robotics, and aviation.
Dr. Manjari Chandran-Ramesh, Partner at Amadeus Capital Partners, highlighted that despite AI’s vast potential, its implementation in industries like banking and insurance remains limited due to concerns over model reliability. “Safe Intelligence can identify fragilities, tackle them, and unlock AI’s true potential across diverse sectors,” she said.
The company, a spin-off from Imperial College London, focuses on providing tools and support for deep validation of machine learning models. This is essential for high-risk sectors, where even minor AI failures can have catastrophic consequences. Safe Intelligence’s proprietary ‘robustification’ process ensures AI models perform reliably in the real world by addressing weak points before they can lead to errors.
In industries like finance, where AI-driven lending decisions are gaining ground, Safe Intelligence’s solutions offer a way to ensure these models can adapt to unexpected inputs without jeopardizing stability or fairness. The company’s platform allows for large-scale formal verification of neural network and decision tree models, with plans to expand support to language models in the future.
The leadership at Safe Intelligence, including founder Professor Alessio Lomuscio, a distinguished expert in AI safety, brings decades of expertise. Lomuscio, who has authored over 200 research papers and holds numerous AI safety patents, is driving the company’s mission to ensure robust and trustworthy AI.
Steven Willmott, the company’s newly appointed CEO, brings a wealth of experience in the tech industry, having co-founded and led 3scale, an API management company acquired by Red Hat. Willmott emphasized the need for better AI validation, stating, “As systems become machine learning-based, we must improve our ability to validate and trust these models, particularly in mission-critical applications.”
Editorial Opinion:
Safe Intelligence’s approach to AI validation and safety couldn’t come at a more critical time. As artificial intelligence continues to evolve and become integrated into key sectors like banking, insurance, and aviation, the risk of AI failures remains a major concern. Safe Intelligence’s robustification process represents a crucial step toward mitigating these risks, ensuring AI models perform not just under ideal conditions but in the unpredictable real world. Their focus on validating AI models for safety is a game-changer, particularly as AI-driven systems become more autonomous and complex.
What makes Safe Intelligence particularly promising is the team behind it. With founder Professor Alessio Lomuscio’s deep academic expertise and CEO Steven Willmott’s entrepreneurial background, the company combines academic rigor with practical, industry-relevant solutions. Their ability to identify and address vulnerabilities in AI systems before they impact real-world applications places Safe Intelligence in a strong position to shape the future of AI in high-stakes environments.
As industries look to embrace AI at scale, it’s companies like Safe Intelligence that will lead the charge in making these technologies trustworthy and secure, helping companies navigate the complexities of machine learning with confidence.
If you need further assistance or have any corrections, please reach out to editor@thetimesmag.com.
