The Centre for Data Ethics and Innovation (CDEI) has published its work programme for its first year. Its main focus will be in two streams; online targeting and bias.
The Centre was created to ensure that the UK keeps up with developments in data-driven technology. As well as the two streams mentioned, it will look at ethical and innovative uses of data-driven technology, and make recommendations to guide future government activity. The way it will work is detailed in the graph below.
It was set-up following a consultation last year and builds on existing work by the Department for Digital, Culture, Media and Sport (DCMS).
Steam 1: online targeting
CDEI defines online targeting as the “the identification of individuals or groups, using data about them (or about others who are like them) and their online behaviours and preferences, and the corresponding delivery of personalised messages, content or services to them online.”
Examples include online advertising, search and news feed optimisation and personalised recommendations.
The review will investigate how and why online targeting happens, the potential benefits and harms to users, and possible frameworks to guide the process in the future. It will be carried out in partnership with DCMS and ICO, as well as public and private sector organisations. An interim report is scheduled for Summer 2019, with final recommendations by the end of the year.
Stream 2: online bias
This stream of work is focused on algorithmic bias in a range of sectors including financial services, local government, recruitment and crime and justice. These sectors have been selected for four reasons:
- There is the potential for the use of algorithmic decision making
- Decisions made in these sectors have significant impact on people’s lives
- There is a risk of algorithms generating or worsening biased decision making
- The potential for algorithms to address any existing bias in decision-making
The review will class areas of bias based on the characteristics protected under the Equality Act 2010, as well as considering digital literacy (following on from the publication of the disinformation and fake news report by DCMS last year). Updates on this stream of work will be published at the same time as the online targeting element detailed above.
It’s an important area, as many examples of bias have already been found. Last year, Amazon was in the news because its recruitment algorithm was biased against women, and policing software used in the US, has led to neighbourhoods being proactively targeted based on the racial characteristics of its residents.
What else will CDEI do?
CDEI will also run workshops with key stakeholders from government, public and private sector and end users. These will look at public trust in data-driven technology and the implications for future regulation, guidance and best practice. The organisation will also respond to current issues as and when they arise, and in 2020 it will publish a State of the Nation Report to bring everything together.
It will also have to work with other organisations to avoid competing in the same space, such as the Office for Artificial Intelligence and the AI Council.
For people working in communications, consistency of governance and agreed ethical approaches will be essential. This will ensure that there is clarity around organisational responsibilities and stakeholder expectations. The success of CDEI will depend on bringing everyone together so that there is an agreed way forward to manage the important issues that technology and data are creating.
Image by Billy Huynh, via Unsplash