Skip down to main content

Consumers missing out on products and offers through tech giants unfair profiling in online advertising, say Oxford legal and AI ethics experts

Published on
27 Apr 2021
Leading experts in ethics and law from the Oxford Internet Institute believe current European Union (EU) legislation is failing to ensure consumers do not miss out on products and offers based on their online behaviour.

Leading experts in ethics and law from the Oxford Internet Institute (OII), University of Oxford, Dr Johann Laux, Postdoctoral Researcher, Professor Sandra Wachter, Associate Professor and Senior Research Fellow and Dr Brent Mittelstadt, Senior Research Fellow believe current European Union (EU) legislation is failing to ensure consumers do not miss out on products and offers based on their online behaviour.

Advertising based on online behaviour (known as Online Behavioural Advertising, or OBA) makes assumptions about peoples’ preferences and interests, based on data in their social media profiles and online browsing histories. While this can help companies better target potential consumers, Google’s and Facebook’s market power in ad-tech gives advertisers further potential to exploit online consumers in their purchasing decisions.  This means that consumers could be paying higher prices than they should or missing out on hearing about products and offers that might actually have more direct relevance to their lives.

In their new paper, ‘Neutralising Online Behavioural Advertising’, forthcoming in Common Market Law Review journal, the Oxford researchers explain how EU consumer law, specifically the Unfair Commercial Practices Directive (UCPD) could be utilised to help mitigate the negative impact of online behavioural advertising on consumer choices.

Dr Johann Laux, researcher at Oxford Internet Institute, University of Oxford, and lead author of the paper said:

“Online behavioural advertising relies on assumptions being made about peoples’ interests and affinities to target consumers based on their online behaviour.  This technology can improve the matching of adverts with consumer preferences, but it also poses risks to consumer welfare through offer discrimination and exploitation of their cognitive errors whilst online.

“The European Union has recently made sensible proposals for new laws to regulate digital platforms holding market power. However, a quicker way to protect consumers today would be to rely on current consumer law and give it an interpretation that accounts for the dominance of a few companies in the ad-tech industry. The legal possibility is there, as we suggest in our new paper.”

The paper highlights that online behavioural advertising, combined with the market power of a small number of dominant technology firms, can lead to consumers seeing less non-personalised adverts which might still be relevant, with consumers missing out on certain products and services.

Laux, Wachter and Mittelstadt make the case that judges and national consumer authorities should consider how the concentration of market power within the ad-tech industry affects the transactional decision making of behaviourally targeted customers.

Adds Laux, “At the moment there are a relatively small number of platforms providing online advertising services who hold significant market power. This is leading to consumer choice, a core component of EU consumer law, being threatened.”

In the paper, the Oxford academics explore how the concept of market power can be incorporated into an assessment of a UCPD violation.  They examine whether targeted advertising and the market power of dominant advertising platforms interact in such a way to encourage consumers to make transactional decisions they otherwise would not have taken if they hadn’t been targeted by online advertising.

The paper sets out how online behavioural advertising can amount to a misleading action and/or a misleading omission according to Articles 6 and 7 of the UCPD.  According to the Oxford analysis, online behavioural advertising can also constitute an aggressive practice under Article 8 of the UCPD.

Drawing on research from economic theory, the researchers put the case for deploying a stricter ‘average consumer test’ under UCPD, if the market for advertising intermediaries is highly concentrated.  The test considers how likely it is that the targeted consumers are seeing an advert that is deliberately exploiting an individuals’ predictable irrationality in their transactional decision making.

The paper recommends how regulators, authorities and courts should use this test as the benchmark by which to assess behaviourally targeted advertising, whenever it singles out an identifiable group of consumers.

Professor Wachter adds, ‘Dominant market actors that engage in the exploitation of consumers’ inferred cognitive dispositions should be under stricter assessments of EU consumer law, particularly in digital markets.  In our paper we’re calling for a case-by-case analysis with a heightened scrutiny of the market environment that online behavioural advertising operates in.  Our approach builds on existing provisions in EU consumer law and argues that the Unfair Commercial Practices Directive can play a constitutive part in mitigating consumer harms created by online advertising practices.”

”In a previous paper, ‘Affinity Profiling and Discrimination by Association in  Online Behavioural Advertising’ published in Berkeley Technology Law Journal, Wachter presented an alternative legal concept to address structural bias in online advertising.

For more information call +44 (0)1865 287 210 or contact

Notes to Editors

About the paper

Dr Laux, Professor Wachter and  Dr. Mittelstadt’s paper  ‘Neutralising Online Behavioural Advertising’, is available as a pre-print and is forthcoming in Common Market Law Review, Volume 58, Issue 3.  This paper is a deliverable of the ‘A right to reasonable inferences in advertising and financial services’ project. This project is financially supported by the Miami Foundation, Luminate Group, the British Academy (grants PF170151 and PF2180114), and the EPSRC via the Alan Turing Institute (grant EP/N510129/1).

Related Topics