Recommender Systems – Legal and Ethical Issues Interdisciplinary Conference
Talks at this conferenceAdd a talk
‘Recommender systems’ are fully or partially automated systems used by an online platform to suggest in its online interface specific information to recipients of the service. We invite paper proposals from all scholars working on the philosophical, ethical, and legal aspects of recommender systems and their environments as well as from scholars working in neighboring, related fields. The conference aims to further the interdisciplinary, transdisciplinary and transnational discussions. Papers will be published in the conference proceedings.
AI-powered recommender systems are now a ubiquitous part of our lives. The next news article we
read, movie we watch, song we listen to, product we buy and social media post we see, are most
likely recommended to us by an algorithm. Due to their global and fast spread in the last decade, both legal and philosophical research has yet to explore implications that recommender systems
have for our lives in a digitalized society. The European Commission, in its Proposal for a Digital
Services Act, defines ‘recommender system’ as fully or partially automated system used by an online platform to suggest in its online interface specific information to recipients of the service. Recommender systems determine relevance for users, oftentimes using personal data and profiles to determine the relative relevance of recommended content. The ever-changing technological landscape requires a detailed philosophical and ethical analysis to both evaluate existing legislation and shape future regulatory measures.
The use of recommender systems gives rise to various philosophical questions. How can relevance be defined? Whom are recommendations benefitting? More fundamentally, how should we understand the concept of a recommendation? This understanding will feed into an ethical framework for recommender systems which mitigates the widespread harms associated with them. These harms include: exposing children to inappropriate content, personal information being inferred by others who see which items get recommended to a particular user, reducing choices
offered to users, and recommending items in a way that is unfair to particular groups of users. Developing a systematic ethical framework for recommender systems will help those designing them to reduce these harms. The respect of ethical guidelines in the framework at design stage can be part of legal instruments, such as a certification. Other possible legal approaches include sector-specific regulation of recommender systems concerning, for example, news or online marketplaces. Sectorspecific (national) paths to the regulation of recommender systems are to be analyzed in comparison to the proposed EU Digital Services Act, which - in its current version - barely scratches the surface of the regulatory need associated with recommender systems. It limits the regulation of recommendation systems to large online platforms, which then have to provide transparency and choice of options for users.
We invite paper proposals from all scholars working on the philosophical, ethical, and legal aspects of recommender systems and their environments as well as from scholars working in neighboring, related fields. The conference aims to further the interdisciplinary, transdisciplinary and transnational discussions. Papers will be published in the conference proceedings.
Possible topics include:
1. Conceptual analysis of ‘recommendation’ and of how ‘relevance’ is interpreted differently across recommender systems
2. How different stakeholders are affected (differently) by specific recommender systems
3. How the design of recommender systems promotes certain values and constrains others
4. Relevance for specific sectors, such as fake news or market places
5. Regulability of recommender systems and its limits
6. Media law(s) and recommender systems, including comparative legal analysis
7. Data protection and recommender systems, including comparative legal analysis
8. Platform regulation and recommender systems, including comparative legal analysis
9. Certification for recommender systems
Abstract submissions are due on July 15th 2021. They should include the following elements:
b. Abstract (800 – 1.000 words), preferably in Microsoft Word Text format (.docx)
c. Author’s name, affiliation and short bibliographical note (in the body of the email)
All submissions shall be sent to Sergio Genovesi ([email protected])
Who is attending?
No one has said they will attend yet.
Will you attend this event?