AI research receives a major boost with £80M national initiative

By analysing trends in patient care records, it might be possible to predict whi
By analysing trends in patient care records, it might be possible to predict which therapies will be most effective for patients with mental health issues. (Credit: Prostock-studio)

The AI for Collective Intelligence hub (AI4CI) is a £12M, five-year collaboration between the University of Bath and six other institutions.

  • Published on Tuesday 6 February 2024
  • Last updated on Tuesday 6 February 2024

Developing next-generation artificial intelligence to tackle complex global problems is the theme of a major new project launching nationwide today, with experts from the University of Bath playing a key role.

The £80M project aims to transform the way AI is both developed and used. Among other things, scientists will research new ways to tackle climate misinformation, ease the flow of traffic in smart cities and deliver healthcare that is tailored to the individual patient.

Another of the project’s missions is to explore how AI networks can help humans make smarter decisions during major health events such as a pandemic. For this strand of the project, experts will reconsider the modelling and analyses undertaken during the Covid-19 pandemic, exploring how new AI approaches could improve centralised policy making and empower individual decisions during a future pandemic.

The Bath component of the project is being led by Dr Theresa Smith in the Department of Mathematical Sciences and Professor Laura Smith in the Department of Psychology. Dr Cangxiong Chen from the Institute for Mathematical Innovation will be the early career researcher lead. The Institute for Digital Security and Behaviour is also involved.

Dr Theresa Smith said: "All of us are impacted by AI - this new project will allow us to collaboratively design new sustainable and responsible approaches that integrate large-scale data streams to support individual and collective decision-making for everyday challenges."

Nine new hubs

Nine new research hubs located across the UK and funded by the Engineering and Physical Sciences Research Council will be involved in the project. Of these nine hubs, three will address the mathematics and computational research that is foundational to AI. The remaining six will explore AI for science, engineering and real-world data, providing the tools needed to accelerate future AI innovations and advance its application in key areas such as healthcare.

University of Bath academics will contribute to a hub exploring real-world data. Named AI for Collective Intelligence (AI4CI), the hub is a collaboration between the Universities of Bath, Bristol, Cardiff, Exeter, Glasgow, and Ulster and University College London (UCL). It is being led by Bristol and funded to the tune of around £12M over five years. Of this, £1.33M has been awarded to Bath.

Collective intelligence describes the process of pooling the information acquired from independent computer systems to create AI networks that offer new, intuitive ways for people to interact with the world.

The AI4CI hub will generate new AI tools that will leverage the collective intelligence distributed across devices and populations of people in order to improve both individual and collective decision making. This already occurs naturally when people share, compare, and filter information, but new AI offers the possibility to harness this kind of collective intelligence at scale, delivering tailored support and guidance.

Professor Laura Smith explained: "We are thrilled to bring our expertise to bear on the real-life applications of AI, to ensure that the solutions the hub delivers are accessible and inclusive for the people they are designed to support.

"AI often addresses challenges that are relatively monolithic in nature, such as: determine the safest route for an autonomous car; translate a document from English to French; analyse a medical image to detect a cancer.

"These kinds of challenge are very important and worthwhile targets for AI research, however, there are challenges that are more ’collective’ in nature, involving a large number of people, and which unfold in real time: these are the ones we want to address."

For example, taking the experiences of patients receiving treatment for anxiety and depression in the NHS. By identifying trends in care records, it might be possible to anticipate which types of therapy will be most effective and deliver bespoke guidance directly to patients and providers.

Dr Theresa Smith said: "The aim here would be to help populations of people with common mental health conditions reach recovery more quickly by identifying patterns in their pooled disease trajectories while preserving their privacy and anonymity."

Key challenges in this case include identifying patterns that are robust, anticipating and understanding changes in a patient’s circumstances, and making useful personalised recommendations, which also ensure privacy, safety, fairness, and trust.

Dr Chen said: "The issue of data privacy is paramount in this project. To maximise the capability of collective intelligence, we’ll need to build AI that has seen vast amounts of data across a variety of real-world scenarios, and this will include sensitive data, such as information about the provider and the data they provide.

"I will be leading research on developing privacy-preserving methods to hide this data that could potentially be uncovered during the training of AI models.