UK Government Offers Up to £400k in Funding to Tackle AI Bias in Healthcare and Real-World Use Cases
UK companies now have the opportunity to apply for government investment of up to £400,000 to develop innovative solutions that address bias and discrimination in AI systems. The Fairness Innovation Challenge, delivered through the Centre for Data Ethics and Innovation, aims to support up to three groundbreaking projects, with each successful bid receiving a funding boost of up to £130,000.
The competition has been launched in anticipation of the world’s first major AI Safety Summit, which will explore ways to manage the risks associated with AI while maximizing its potential benefits for the British people.
The Department for Science, Innovation, and Technology’s challenge seeks to nurture the creation of new approaches that prioritize fairness in the development of AI models. By encouraging participants to incorporate a wider social context into their models from the outset, the challenge aims to address the threats of bias and discrimination.
Fairness is one of the key principles for AI outlined in the UK Government’s AI Regulation White Paper. While AI has the potential to drive economic growth and improve public services, it also poses risks that must be addressed. In the healthcare sector, for example, AI is already being used by the NHS to aid in the identification of breast cancer cases. It also holds promise in developing new treatments and tackling global challenges like climate change. However, these opportunities can only be fully realized if bias and discrimination are effectively tackled.
Viscount Camrose, Minister for AI, emphasized the need to address the risks associated with AI in order to fully harness its benefits. By ensuring that AI models do not reflect biases present in society, AI can become safer, fairer, and more trustworthy. Additionally, a UK-led approach is being promoted to align with the country’s specific laws and regulations.
The Fairness Innovation Challenge will focus on two areas. Firstly, a partnership with King’s College London will allow participants from the UK’s AI sector to work on addressing potential bias in a generative AI model. This model, developed in collaboration with Health Data Research UK and the NHS AI Lab, leverages anonymized patient records to predict possible health outcomes.
Secondly, the challenge welcomes proposals for new solutions that tackle discrimination in various models and areas, such as fraud prevention, law enforcement AI tools, and fair recruitment systems.
Challenges faced by companies in addressing AI bias include insufficient access to demographic data and ensuring compliance with legal requirements. To assist participants, the Centre for Data Ethics and Innovation is collaborating with the Information Commissioner’s Office and the Equality and Human Rights Commission to provide guidance and expertise on data protection, equality legislation, and mitigating bias in AI development.
The Fairness Innovation Challenge will also offer assistance in applying assurance techniques to AI systems to achieve fairer outcomes. Assurance techniques involve verifying and ensuring that systems meet certain standards, including fairness.
Baroness Kishwer Falkner, Chairwoman of the Equality and Human Rights Commission, stressed the importance of careful design and regulation to prevent AI systems from disadvantaging protected groups. Both tech developers and public authorities have a responsibility to ensure that AI systems do not discriminate and comply with equality legislation.
The submissions for the Fairness Innovation Challenge will close at 11 am on December 13, 2023, and successful applicants will be notified on January 30, 2024.
In conclusion, the UK Government’s funding initiative seeks to address bias and discrimination in AI systems and promote fairness. By supporting innovative solutions, the government aims to maximize the benefits of AI while mitigating potential risks. The challenge encourages the incorporation of a wider social context in the development of AI models, ensuring they align with UK laws and regulations. The initiative will focus on healthcare and real-world use cases, and successful applicants will receive funding of up to £400,000.