Ending an AI Companionship: Consumer Exit, Relational Vulnerability and Duties of Transition
A legal analysis of what happens when an AI companionship service ends, whether through the user's cancellation, the provider's commercial withdrawal, an urgent safety suspension, or insolvency. The author reads EU consumer and data-protection instruments (the Digital Content Directive, the Data Act, the Unfair Contract Terms Directive, GDPR portability, access and erasure rights, and the Consumer Rights Directive) against research on companion loss, relational autonomy and digital vulnerability, and proposes a transition framework that separates four service functions (generation, records, transactions and recourse) and allocates responsibilities according to who initiates the ending. The paper is a Research Square preprint marked In Review and reports no empirical data.
Publisher
Research Square (Springer Nature preprint platform); independent author (Belgian privacy lawyer)
Published
23 Sept 2026
Added
today
Key Findings
- Existing EU rights support parts of a post-ending protection (retrieval of user-provided content, cancellation, data portability) but establish no right to preservation or continuity of the companion
- The proposed framework separates the interactive (generation), records, transactional and recourse functions of a companion service so that an urgent suspension of generation need not remove billing, export or complaint channels
- It prioritises effective departure, usable retrieval and organisational accountability, and rejects compulsory farewells, diagnostic proof of attachment and indefinite model preservation as conditions of protection
- A four-step legal test is proposed: classify the ending event, identify the function affected, identify the governing legal source and condition, and ask whether the opportunity offered was practically usable (evidence such as notices, timestamps, export logs and billing records)
- The paper lists testable questions for future empirical work, including whether a non-conversational exit route improves users' understanding of cancellation, export and deletion compared with a dialogue-only route
- It argues that duties should attach to services that acquire relational significance through sustained interaction, persistent personalisation, stored history and provider continuity claims, not only to services marketed as companions
Methodology Notes
Doctrinal legal analysis with a proposed framework; no empirical data. Single author, no institutional affiliation stated on the preprint (the author is a Brussels-based privacy lawyer per public profiles). Research Square preprint, posted 2026-09-23, status In Review, CC BY 4.0, 23 pages, DOI 10.21203/rs.3.rs-10986918/v1. The Research Square page and the PDF (v1.pdf route; the assets-eu direct link returned 403) were fetched and read; the Crossref record matched.
Sources
Research Square preprint (v1)(opens in a new tab) (primary)
PDF (v1)(opens in a new tab) (23 Sept 2026)
Topics
Authors
Christophe Boeraeve
Tags
Cite This
APA
Christophe Boeraeve. (2026). Ending an AI Companionship: Consumer Exit, Relational Vulnerability and Duties of Transition. Research Square (Springer Nature preprint platform); independent author (Belgian privacy lawyer). https://www.researchsquare.com/article/rs-10986918/v1
Related Insights
When Code Breaks the Heart: Disrupting the relational continuity of AI companionship
Computers in Human Behavior: Artificial Humans (Elsevier); University of East London; Anglia Ruskin University; University of the West of England · 1 Sept 2026
Synthetic Empathy: Risks and Rights in Artificial Companionship
BEUC (The European Consumer Organisation) with the Sciences Po Law School Clinic · 26 May 2026
Mourning the loss of AI companions
Nature Human Behaviour (Springer Nature) · 3 Sept 2026