Close
Search
A content on our website (by Google)
Close
Search
A content on our website (by Google)

Statement on the use of artificial intelligence at ULiège Library

Generative artificial intelligence (AI) is now present in services and systems used at ULiège Library, whether through features integrated by our suppliers into our library systems or through solutions that we are exploring to enhance our services. This statement aims to provide all our users (members of the university community and external users), as well as our partners, with transparent information about how AI is used, the principles that guide our choices, and the uses that we deliberately exclude.

Our approach is grounded in the fundamental values of academic libraries, and of ULiège Library in particular: equitable access to information, respect for privacy, intellectual integrity, and service to the academic community. It is also aligned with the University of Liège’s principles for responsible digital governance and its charter on the use of generative artificial intelligence in academic work.

ULiège Library recognises that the ethical questions surrounding AI are complex, often contested, and evolving alongside technologies and their contexts of use. We therefore favour a reflective and iterative approach over fixed positions, drawing in particular on recognised frameworks and reference documents. This reflection also encompasses issues relating to the use of heritage and unique collections in the context of training AI systems, in line with the Statement of Shared Practice: AI Training Requests and Unique Cultural Collections (2026).

This document is intended to evolve. As technologies and the needs of our community develop, the practices described here will be reviewed and updated. Any changes will be communicated openly.

Guiding principles

All uses of AI at ULiège Library are guided by the following principles.

Human oversight Transparency Privacy and data protection Fairness and non-discrimination Integrity and reliability AI literacy Staff training and skills development Continuous evaluation

Examples of AI use at ULiège Library

In the interests of transparency, the examples below illustrate some of the AI functionalities already integrated into our user-facing services and systems.

  • An AI metadata assistant that suggests enrichments to bibliographic records.
  • The automatic generation of textual descriptions for images and digitised documents, primarily those of heritage value, to improve their accessibility and discoverability.
  • An AI research assistant that allows users to query documentary resources in natural language and receive concise answers accompanied by verifiable references, drawn from an index of more than 5 billion identified and verifiable academic records.
  • A virtual assistant integrated into the online chat service and trained on internal documents. Available in particular during evenings and at weekends, it can answer users’ questions about ULiège Library’s services and collections.
  • The analysis of usage data for physical and electronic collections to inform acquisition and collection management decisions, based on aggregated and anonymised data, and to help improve the quality of our collection data.
  • The use of AI-based programming assistance tools to support internal software development, in accordance with best practices for code quality, security and data protection.
  • The analysis of non-sensitive datasets to identify trends, produce indicators, and facilitate the preparation of activity and management reports.

What we do not do with AI

The following commitments are firm. They define the boundaries that ULiège Library sets for its use of AI.

User surveillance and profiling Processing of sensitive or confidential data Communication

On vendor-provided tools

Most of the systems used at ULiège Library are third-party products that incorporate, and continue to progressively incorporate, AI functionalities. These functionalities are often offered as optional modules rather than as integral and inseparable components of the systems.

ULiège Library:

  • evaluates these functionalities before enabling them, whenever possible;
  • engages with vendors on data practices, algorithmic transparency and GDPR compliance;
  • reserves the right to disable AI functionalities that prove unsatisfactory, opaque or inconsistent with our values;
  • informs the community about significant developments in the tools it provides.

Trust in established and reputable vendors does not remove the need for active oversight. The relevant staff discuss developments with vendor representatives and test new functionalities before they are put into production, whenever possible.

Governance and responsibilities

Responsibility for implementing this statement lies with the ULiège Library Directorate, in coordination with the relevant services of the University of Liège. The Library’s practices are reviewed internally on a regular basis, and at least annually.

Any questions, comments or concerns regarding the use of AI at ULiège Library may be addressed to the Library Directorate or to the University’s Data Protection Officer.

References and resources

updated on 9/7/26

Share this page

cookieImage