About Algorithms and Beyond is a project supporting the multidisciplinary development of a cohort of five interns that will work at Horizon Digital Economy Research on algorithm-related topics. The vision being to immerse the cohort on the multidisciplinary ethos of Horizon, whilst working collaboratively with other schools at the University of Nottingham, i.e. Law and Culture, Film and Media studies.
Each intern will work both collaboratively within the cohort to gain insight about the many overlapping implications that algorithms have, as well as independently, on the following research projects.
Implementing an Open Educational Resource (OER) designed to promote awareness on algorithm fairness and bias among children and young people (CYP), led by Elvira Perez Vallejos.
This project drew upon the results from a previous feasibility study conducted to evaluate a similar OER designed to promote digital rights among children and young persons (CYP): http://casma.wp.horizon.ac.uk/casma-projects/5rights-youth-juries/oer and focused on developing and implementing user’s recommendations into a new Online Educational Resource (OER) designed to promote awareness on algorithm fairness and bias among CYP. The aim of the project was to ensure the OER is more equitable, accessible and inclusive for users.
Enabling Social Impact Assessments for Algorithmic Systems, led by Lachlan Urquhart (working with Dimitrios Darzentas in the Mixed Reality Lab).
This project will involve the design of prototype ideation cards and disseminate an initial survey, drawing on perspectives from IT law, digital ethics and human computer interaction. We will design prototype decks of ideation cards to support reflection by designers on end-users’ privacy, security, legal rights and ethical interests when using technologies underpinned by algorithms, like the Internet of Things. Our methodological process for using the cards will draw on three frameworks: value sensitive design, human data interaction and ethical/privacy/social impact assessments. We will be working alongside colleagues from the Mixed Reality Lab on the Magellan Project who are building a platform that uses computer vision and augmented reality to track how the cards are used. A survey will also be designed and disseminated to Horizon partners, which will build up understanding of the motivations and challenges partners face in compliance with privacy, security, legal and ethical standards.
Mapping the capabilities of person profiling algorithms, led by Ansgar Koene.
This project is a literature study into the data aggregation and analysis methods used by industry, or proposed in academic literature, for creating person profiles used in algorithmic personalization of media content, advertising and search/recommender system. Focus will be on two aspects: 1. Privacy impact, e.g. ability to infer aspects about a person that were not explicitly in the input data; 2. Susceptibility to bias. This work will contribute to the UnBias project, will feed directly into our work on the P7003 IEEE standard on Algorithm Bias Considerations, and also our engagement with the UN Special Rapporteur on Privacy.
The Role of Consumer Law in the Governance of Targeted Advertising Algorithms, led by Richard Hyde.
Consumer law was written with traditional forms of advertising in mind. It is not well adopted to situations where algorithms are used to individually target consumers, and therefore much of the work in regulating targeted advertising is done by data protection and privacy law. However, consumer law (such as the Unfair commercial Practices Directive and the Unfair Terms Directive) can and should play an important role in the governance of targeted advertising.
This project builds on work that examines online targeted advertising by considering, first, the ways that data from beyond browsers may be used by algorithms to target advertising and, second, the challenges and opportunities for governance of such algorithms and their advertising outputs by consumer law and beyond.
Project update – December 2017
Over a productive four month period the Consumer Law Team have developed a theoretical framework to explain the infringement of consumer rights that can result from commercial practices that utilise algorithms.
By utilising the framework provided by the United Nations Guidelines for Consumer Protection, the team have identified the right to free choice, the right to information and the right to privacy as the consumer rights that are engaged by algorithmic commercial practices.
In order to assess whether algorithmic commercial practices posed a risk to consumer rights we carried out a literature review, examining offline commercial practices, online commercial practices and algorithmic commercial practices, assessing whether these practices were low or high infringements of the three rights. We then placed the different practices within an octant diagram, with the practices most in need of regulation in the octant representing high infringement of each of the three rights. We discovered two practices in this octant, subliminal advertising, and emotional targeted advertising.
Subliminal advertising is already banned in most jurisdictions. Therefore, our project focuses on the regulation of emotional targeted advertising, assessing the current regulatory environment, in particular examining the GDPR, the ePrivacy Directive and the Unfair Commercial Practice Directive.
The project concludes that this regulatory framework does not go far enough in regulating this area, particularly with regard to emotionally vulnerable consumers. We will therefore propose regulatory changes to ensure that consumer’s rights are not infringed by emotional targeted advertising.”