Two suppliers who met through ACE’s Vivace community have been awarded a £100,000 grant from the Government’s AI Fairness Innovation Challenge to develop novel solutions for addressing unfair bias and discrimination in artificial intelligence systems.
The Challenge, managed by the Department for Science, Innovation and Technology (DSIT) and delivered by Innovate UK, was created to fund new ways to address statistical, human and structural bias and discrimination in AI systems.
Coefficient and Bays Consulting, who have known each other since 2022 and worked together on projects inside and outside ACE, designed a proposal to highlight potential harms whilst pushing for safer and responsible AI development. The solution adopts a socio-technical approach to reduce unfair bias in CV screening algorithms and applicant tracking systems used in the recruitment sector.
Deliverables for the project, which has been underway since May, will include:
- an extensive market evaluation report of AI-enhanced CV screening tools in applicant tracking systems.
- a framework for measuring unfair bias from HR systems using algorithmic decision-making.
- an open-source bias mitigation toolkit for machine learning engineers.
- an outreach and knowledge transfer programme of newsletters and events.
- design and development of technical service for AI safety accreditation and assurance.
John Sandall, CEO of Coefficient, said: "Coefficient are building on the ideas outlined in ‘Weapons of Math Destruction’, focusing on the intricate problem of algorithmic bias—a result of prejudiced training data that can perpetuate systemic inequalities and create feedback loops.
"This issue's complexity is compounded by the "black box" nature of AI systems in recruiting software, which often lack transparency and explainability."
Dr Sophie Carr, CEO of Bays Consulting, said: "This project is not just about providing a technological solution; it's about fostering an equitable recruitment environment that champions diversity and fairness where 86% of companies are incorporating AI into their hiring processes.
"Across the board, the UK's approach to algorithmic auditing is governmental rather than auditable, leaving much room for improvement and a pressing need for standardised oversight.
"We aim to set a new standard in the recruitment landscape, ensuring our AI tools serve to support unbiased decision-making. It’s wonderful to have this opportunity to extend the knowledge we gained in CV screening through commissions in ACE and build this into a solution that helps both companies involved grow."
The Coefficient/Bays bid and the three other winners, the Open University, the Alan Turing Institute and King’s College London, were selected based on potential impact and innovation by expert assessors chosen by DSIT and Innovate UK. The proposed AI regulatory principles, including fairness, are set out in the UK government’s white paper, “A pro-innovation approach to AI regulation”.
Subscribe to the team’s new community outreach newsletter, ‘Fairness Tales’, to stay informed on progress and insights.
Leave a comment