What is the MuseumWeek AI Academy?+
An initial training designed for museums and cultural organizations: two short tracks to decide on and frame a use, or to test it in your profession. A common foundation covers AI limits, reliability, data, rights, and human verification. Institutional projects, including creating a charter, can also be supported separately.
Who is it for?+
Leadership, managers, AI leads, and subject-matter specialists can follow either of the two tracks. These four profiles help adapt examples and exercises but do not correspond to four separate catalogs. The AI lead often carries this mission alongside another job. The training is open with your institution or individually; it requires neither coding skills nor an existing AI policy.
Is training staff in AI mandatory?+
In the European Union, Article 4 of the AI Act requires providers and deployers to take measures to support the development of AI literacy among people who use these systems on their behalf, taking into account their knowledge and usage context. It does not require a specific training course, nor a set level for each person. Training adapted to roles and uses can contribute to this effort. The Academy offers guidelines, exercises, and materials tailored to profiles, without replacing the assessment of each organization's specific needs.
What distinguishes it from other AI trainings?+
The training starts from a case in the museum sector and combines hands-on practice, critical judgment, and usage rules. It leads to a first verifiable result, without promising to deploy a tool in a few hours. Questions of charter, pilot, integration, and evaluation can be the subject of tailored support, independently of the training. The UNESCO-ICOM 2026 survey calls precisely for practical, museum-specific training combining technical skills and ethical reflection.
And our collections, our data, our rights?+
This is the Govern step, and it comes second, not last. Our survey shows the issue is first and foremost ethical: 52% of professionals surveyed cite ethical concerns among their organization's obstacles, twice as many as legal questions (27%). In practice, the difficulties encountered are very concrete: among museums already using AI, the UNESCO-ICOM 2026 survey ranks the accuracy of generated content first (79%), then copyright and intellectual property (63%), then data protection (61%). We therefore begin with what the institution allows itself and on what grounds, before specifying which data may enter the tools. During the training, participants apply these safeguards to the uses and trials they work on; creating or deploying an assistant falls under separate support. Scientific and editorial validation always remains human: no format proposes doing without it. Indeed, 81% of museums using AI report always having a person review what it produces before using or publishing it.
And the environmental impact of AI?+
Using AI consumes energy: this question is legitimate, especially for institutions committed to climate issues. It falls under the Govern step. It involves deciding which tasks justify an AI tool and which do not, preferring the simplest tool that suffices, and embedding these choices in the charter. In the UNESCO-ICOM 2026 survey, 47% of museums using AI cite its environmental impact among the challenges encountered; only 29% declare taking it into account in their use. The goal is not to use AI everywhere, but where it truly serves, according to the principle of necessity also advocated by the Metropolitan Museum of Art in its survey responses.
Will AI replace jobs in our teams?+
Two out of three professionals surveyed in 2025 want to better understand what AI changes in cultural work and professions: the question arises and is addressed openly. On the institutional side, 35% of museums using AI cite its impact on professions among the challenges encountered; in their open responses, museums most often describe it as support for professional judgment, not as a substitute (UNESCO-ICOM 2026 survey). Training starts from tasks, not positions: what a tool can prepare, and what remains human judgment, such as scientific validation, editorial choice, or the relationship with audiences. Every automated task retains a human control point. The goal is to give time back to the core profession, not to reduce teams.
What are the formats?+
Two tracks are offered: “Decide and frame” (4 hours) or “Practice in your profession” (6 hours). Each includes a common foundation and application time. Sessions take place online, by videoconference. Individually, you may request a spot in a group session with professionals from other institutions; for an institution, we adapt the proposal to the team and can plan follow-up after the session. Its pace and objectives are defined with the institution. Finalizing a charter or deploying a tool requires separate project support.
What skills do participants develop?+
In both tracks, participants learn to identify a relevant use, give useful instructions, test a result, and decide when not to use AI. They also practice verifying facts, sources, rights, and data. The “Decide and frame” track applies these benchmarks to the institution's choices; “Practice in your profession” applies them to a real task or, if needed, to a provided example. The immediate results remain those announced for each track: a decision grid and a draft of rules, or a working method and a quality checklist.
How much does it cost?+
For an institution, we tailor the proposal to the museum’s size and budget, the number of people to train, the format, and any follow-up requested. For individual participants, the fee depends on how many people take part in the session. In both cases, we share the terms before you register.
Is coding necessary?+
No technical level is required. Knowing a real process from your institution and bringing a use case or some non-sensitive documents can help, but it is not mandatory: exercises can also rely on provided examples. The AI lead is often someone from communication, documentation, public engagement, or administration who carries this mission in addition to their role.
In which languages are sessions held?+
Sessions take place online, by video conference, in French or English only.
How do we start?+
In “Choose a plan,” indicate whether you are participating individually or with your institution, then choose one of the two tracks. The configurator prepares a summary to send by email; it does not constitute final registration. An institution receives a proposal tailored to its team and resources. Individually, you may request a free 20-minute call to clarify your needs before requesting a spot in an online group session.
