CDEI review considers algorithmic bias
The Centre for Data Ethics and Innovation (CDEI) is an independent advisory body, led by a board of experts, set up and tasked by the UK Government to investigate and advise on how to maximise the benefits of data driven technologies. The UK Government is required to consider and respond publicly to CDEI recommendations. CDEI recently launched a review into algorithmic bias. Following a call for evidence in the spring the CDEI has now published its interim reports on this topic.
Background to the review
The CDEI has identified bias in algorithmic decision making as an immediate concern, undermining the effectiveness and reliability of machine learning technology noting that:
If designed well, [algorithms] can reduce human bias in decision-making processes. However, as the volume and variety of data used to inform decisions increases, and the algorithms used to interpret the data become more complex, concerns are growing that without proper oversight, algorithms risk entrenching and potentially worsening bias
CDEI’s review on this topic has focused on four sectors ((i) policing, (ii) financial services, (iii) recruitment and (iv) local government).
CDEI Interim Findings
CDEI’s interim report considers the following questions:
Data: Do organisations and regulators have access to the data they require to adequately identify and mitigate bias?
In their interim findings CEDI note that:
While data itself is often the source of bias, it is also a core element of tackling the issue…. tension between the need to create algorithms which are blind to protected characteristics, while also checking for bias against those same characteristics, creates a challenge for organisations seeking to use data responsibly.
Further guidance for firms in this area will therefore be welcome.
Tools and Techniques: What statistical and technical solutions are available now or will be required in the future to identify and mitigate bias and which represent best practice?
CDEI’s early work suggests that new approaches to identifying and mitigating bias are required. They note that
there is limited understanding of the full range of tools and approaches available (current and potential) and what constitutes best practice. This makes it difficult for organisations that want to mitigate bias in their decision-making processes to know how to proceed and which tools and techniques they should use
Governance: Who should be responsible for governing, auditing and assuring these algorithmic decision making systems?
CDEI notes that:
Decision-makers are likely to face significant trade-offs, for example, between different kinds of fairness and between fairness and accuracy….There is currently limited guidance and a lack of consensus about how to make these choices or even how to have constructive and open conversations about them. These choices are likely to be highly context specific and as such, the way they are made, governed and audited will need to be considered on a sector by sector basis
CDEI intended to consider these issues in more detail on a sector by sector basis as it develops its final recommendations.
CDEI reports that their next steps over the next eight months will include:
- a summary of responses received later in the summer.
- a draft Policing Code of Practice in the autumn for consultation (to be finalised in early 2020)
- continued engagement with financial services and recruitment sector stakeholders and research in these areas
- a final report with recommendations to the Government in March 2020.
Click here to access the interim review