Follow us

UK – DATA & NEW TECHNOLOGY

Following the Centre for Data Ethics and Innovation’s (“CDEI”) review into bias and algorithmic decision-making published on 27 November 2020, the CDEI and the Recruitment and Employment Confederation (“REC”) have published a set of jointly developed guidance to provide practical considerations for the use of data-driven tools (e.g. big data analysis, automation systems, machine learning and other forms of AI), transparency considerations, and commentary on how such tools can add value for organisations (the "Guidance").

Key date(s)

  • 27 November 2020 – The CDEI publishes a review on bias in algorithmic decision-making (the "Review").
  • 11 Nov 2021 – The All Party Parliamentary Group for the Future of Work launches a new report termed 'The New Frontier: Artificial Intelligence at Work' (the "Report").
  • 6 December 2021 – CDEI and REC publish the Guidance.

Status

  • The Guidance published by the CDEI and the REC has been produced in response to calls for practical, sector-specific guidance on the use of data-driven tools for the UK recruitment sector, in particular following the Covid-19 pandemic where the use of digital tools in the workplace surged. In addition to enabling communications such as interviews and sourcing, digital tools, namely data-driven tools, are being used more and more to render recruitment processes more efficient. The Guidance aims to equip recruiters with the knowledge they need to buy and use data-driven tools confidently and critically.
  • The Report found that uncertainty around AI developers' accountability is the main reason for a lack of trust placed in AI tools used in the workplace. The Review also highlighted further issues, including systematic skews in AI-led decision-making and the slipping of protected characteristics under algorithmic radars.
  • According to the Review, 'fairness' is a key issue facing the use of data-driven tools in recruitment. In view of this, the Review recommends that recruiters update their guidance on the application of the Equality Act 2010 to discrimination associated with the use of algorithms. Recruiters should review data-driven tools with the Act in mind, namely bias which leads to unfair treatment based on protected characteristics. However, the Review also warns that the scope of fairness is not limited to the Equality Act. Algorithmic bias may also be non-discriminatory and nevertheless lead to unfair outcomes. The Review states that fair decisions must be non-arbitrary, reasonable, consider equality implications, and respect the circumstances and personal agency of the individuals concerned. Therefore, developers must decide the concepts of fairness that apply to their context and ensure their data-driven tools can deliver. Whilst the Review does not provide specific information on how to achieve this, recruiters should anticipate the need to question developers on their tools' fairness parameters before purchasing.
  • The Guidance is broken down into three main sections and various sub-sections and advises that, as purchasers of digital tools, recruitment companies are clear in their intentions before purchasing such tools, and actively seek information and verify compliance during the purchasing stage. The Guidance also views pilot tests and outcome monitoring as being of equal importance after purchasing and transparency is pinpointed as an essential component for data protection and promoting trust among employees and candidates alike.

 What it hopes to achieve 

  • The Guidance builds on the Report and the Review which identified the vulnerabilities algorithms possess in relation to biases embedded in data, and obscuring individuals' and organisations' accountability in recruitment processes. The Guidance therefore seeks to empower recruitment companies with the knowledge and processes to make smart digital tool purchases and monitor these assets critically and carefully. This is with the aim of reducing the risk of algorithmic biases creating unfair outcomes for candidates at the recruitment stage.
  • In particular, the Guidance names targeted advertising, headhunting software, qualifying screen tools, CV matching, voice and face recognition and offer-predicting software as tools used in the recruitment sector which could present data risks. The Guidance aims to provide recruitment companies with sector-specific advice so that they can implement practical processes and monitoring with clear direction and confidence.

Who does it impact?

  • The Guidance will be of particular importance to UK recruiters, to whom it will provide detailed practical navigation of digitalised processes. Although not binding, it will help recruiters mitigate the risks of failing to comply with data protection laws and the Equality Act.
  • Data-driven tool developers will also be significantly impacted, as the Guidance will likely cause a shift in standardised interactions with tool purchasers. Recruiters may carry out more stringent compliance checks, and developers may find purchasers approach their products with a higher degree of criticality when able to informedly question processes. Developers could find themselves being required to provide deeper understanding of data-driven tools' decision-making processes.
  • As a result of the Guidance's impact on recruiters and developers, job candidates may notice a change in practices when searching, applying and interviewing for jobs. More focus will be directed towards candidate feedback, and transparency regarding data-usage will be engaged to promote trust.

Key points 

  1. Discrimination and fairness
    • The Guidance aims to promote processes which will not only help recruiters develop robust processes to reduce embedded discriminatory biases projected by data-driven tools but also to encourage fairness, which necessitates active engagement with candidates belonging to minority groups and those with protected characteristics.
  1. Data, transparency and due process
    • The Guidance also advises recruiters on effective pre-purchase checks to ensure purchased data-driven tools comply with relevant legislation, including the Equalities Act and applicable data protection legislation.
  1. Various stage monitoring in the recruitment process
    • The Guidance is split into various sections which advise on specific scenarios, including before purchasing, during purchasing, before use, and during and after use. The Guidance highlights that due diligence at every stage of engagement with recruitment software will be the most effective way to mitigate risks.
  1. Holistic approach
    • By combining the law, standards and principles to construct this Guidance, the CDEI and the REC have looked to provide recruiters with a holistic approach to ensure fair use of data-driven tools in this sector can be practically and effectively attained.


Links

CDEI Review in bias in algorithmic decision-making

The New Frontier: Artificial Intelligence at Work

Data-driven tools in recruitment guidance by the CDEI and REC


Related developments

CDEI report on AI to adjust misinformation on social media platforms

 

 

 

This blog post provides an overview of a key recent or upcoming development in digital regulation in the UK or EU as part of our horizon scanning timeline which can be found below.

Contacts

VIEW DIGITAL AND REGULATION TIMELINE  + 

Hayley Brady photo

Hayley Brady

Partner, Head of Media and Digital, UK, London

Hayley Brady
James Balfour photo

James Balfour

Senior Associate, London

James Balfour

Related categories

Key contacts

Hayley Brady photo

Hayley Brady

Partner, Head of Media and Digital, UK, London

Hayley Brady
James Balfour photo

James Balfour

Senior Associate, London

James Balfour
Hayley Brady James Balfour