Stanhope sequred £2.3 mn in funding to develope of 'human-like' AI

The neuroscience-driven AI startup Stanhope has sequred £2.3 mn in funding to enhance the development of ‘human-like’ Artificial Intelligence that models the brain’s method of processing information.

This investment round included contributions from the UCL Tech Fund, Moonfire Ventures, Rockmount Capital, and several angel investors.

Founded in 2021 in London, deep-tech startup team of engineers and scientists lead responsible development programs that build on academic partnerships within the UK.

Stanhope was co-founded by a distinguished team including AI and Neurology Professors Rosalyn Moran and Karl Friston, along with Dr. Biswa Sengupta, a former professional at JP Morgan Chase.

Stanhope was co-founded by a distinguished team including AI and Neurology Professors Rosalyn Moran and Karl Friston, along with Dr. Biswa Sengupta, a former professional at JP Morgan Chase

The intellectual groundwork of Stanhope’s AI models is heavily based on the Free Energy Principle, a theory primarily developed by Professor Karl Friston.

Stanhope’s AI models are based on Professor Friston’s Free Energy Principle, which proposes that the human brain predicts its environment’s needs, minimizing the reliance on extensive sensory data analysis (see how Generative Artificial Intelligence Provides Opportunities for Insurers).

We are bringing Active Inference to market as a novel framework for autonomous decision-making. To contend with the challenges of contemporary AIs, our distinct approach aims to offer true intelligence with critical human explainability and reliability.

This funding round represents strong confidence in applying neuroscience to overcome common AI challenges such as hallucination, lack of adaptability in new settings, and high energy demands.

Stanhope is applying its ‘active inference’ technology in delivery drones, which navigate by analyzing sensor input instead of relying on maps or GPS.

Stanhope sequred £2.3 mn to develope of 'human-like' AI

These drones, along with other autonomous machines, are in use by entities like Germany’s Federal Agency for Disruptive Innovation and the Royal Navy (see How Technologies Will Shape the Future of Autonomous Mobility).

They are undergoing tests indoors to refine their capabilities ahead of more complex outdoor trials set for this summer.

The funds will support further development of its ‘agentic’ models and their practical applications.

AI startups may be some of the hottest investments right now but few have the calibre and deep scientific and technical know-how as the Stanhope AI team

David Grimm, a partner at the UCL Technology Fund

The Stanhope’s blend of neuroscience and advanced AI not only positions them at the forefront of the field but also equips them to tackle some of AI’s most pressing issues.

Stanhope is a neuroscience-driven AI startup that is pioneering the development of artificial intelligence models that simulate the human brain’s process of gathering and processing information.

The technology developed by Stanhope employs what is known as ‘active inference,’ a method where devices such as delivery drones navigate and react to their surroundings not through conventional GPS or map data, but through real-time sensor input analysis.

This approach allows for more efficient processing and potentially less reliance on extensive hardware, which can reduce energy costs and improve versatility in new or changing environments.

The funding Stanhope has secured will be directed towards further research into ‘agentic’ models—a term that refers to AI systems capable of acting with a degree of independence—and the practical application of these models in various fields.

These advancements are aimed at addressing and overcoming some of the most challenging issues facing AI today, such as hallucinations in machine perception, adaptability in new environments, and the high costs of energy associated with running advanced AI systems.

Peter Sonner   by Peter Sonner