UKRI pledges £50m to develop ‘trustworthy’ AI

UK Research and Innovation (UKRI) has announced funding of £50m for the development of ‘trustworthy and secure’ AI that can be applied to societal challenges.

Adobe Stock

The bulk of this money - £31m – will go to Responsible AI UK, a consortium led by the University of Southampton whose goal is to foster the safe and accountable use of AI. Headed by Southampton’s Professor Gopal Ramchurn, the consortium will help people understand what responsible and trustworthy AI is, how to develop it and build it into existing systems, as well as the impacts it will have on society.

 “Trustworthy AI tends to be looked at from a very technical perspective – ie it is tested and validated in well-defined settings,” said Ramchurn, a Professor of Artificial Intelligence at the University of Southampton. “However, that doesn’t mean it will be trusted by the public, government, and industry.

 “AI tends to be looked at by the tech community as AI that has been thoroughly tested. It can be AI that is trustworthy by the technical functionality of the application and the particular closed environments it has been tested in, but it is not trusted because maybe it uses personal data, you know, uses your personal data in ways that you would not want it to do.”

UKRI has also awarded two new Turing AI World Leading Researcher Fellowships, to Professor Michael Bronstein and Professor Alison Noble, both based at the University of Oxford.

Kedar Pandya, executive director, Cross-Council Programmes at the Engineering and Physical Sciences Research Council (EPSRC), said: “The UK’s expertise in the field of AI is a major asset to the country and will help develop the science and technology that will shape the fabric of many areas of our lives.

“That is why UKRI is continuing to invest in the people and organisations that will have wide-ranging benefit. For this to be successful we must invest in research and systems in which we can have trust and confidence, and ensure these considerations are integrated in all aspects of the work as it progresses.”

In addition, £2 million will be awarded to 42 projects to carry out feasibility studies in businesses as part of the BridgeAI programme, aiming to speed up the adoption of trusted and responsible AI and Machine Learning (ML) technologies. The projects will look at developing a range of tools to facilitate assessment of AI technologies, with successful ones going on to receive a share of an additional £19 million for further development.