234 incidents on record · 2026HeadlightsIncident reports by Ellie Harris · Melbourne
5 new this weekLibrary last updated 14 October 2026
← The incident library
HD-INC-197
Workplace relations · France · 2025 · AI tools were made available to a newsroom without the prior consultation of the works council that French labour law required

A Paris Court Suspended a Publisher's Newsroom AI Tools Pending Works Council Consultation

By Ellie Harris · Filed DIGI on the homepage of the group intranet by 12 May 2025; works council proceedings issued 3 June 2025

Alleged: Groupe Moniteur developed or deployed the AI system implicated in this incident. Details are drawn from public reports; parties are presumed innocent of any wrongdoing not established by an official finding.

A Paris Court Suspended a Publisher's Newsroom AI Tools Pending Works Council Consultation

What happened

The Paris Court of Appeal reported in its judgment of 21 May 2026 that Groupe Moniteur, a French publisher of specialist press titles for public sector actors and construction and public works professionals, had introduced AI tools into its newsroom without first consulting its works council. The court confirmed an earlier order requiring the company to suspend its use of the tools until the consultation was completed.

A works council is a group of employee representatives that an employer must consult about certain decisions affecting staff. Under French labour law, as the judgment sets out, an employer must inform and consult the works council before introducing a new technology that is liable to affect employees’ working conditions.

According to the judgment, Groupe Moniteur had introduced a charter that authorised and governed how employees could use AI tools, including ChatGPT. The company had also developed its own AI drafting assistant, DIGI, which could transcribe audio and video, summarise and correct text, reword articles and suggest headlines. By 12 May 2025, DIGI was available to employees through the company’s intranet.

The judgment reported that the works council took the company to court in June 2025 because it had not been consulted before the tools were introduced. In July, a judge ordered the company to suspend its use of AI tools until the consultation was completed and made a provisional award of 5,000 euros in damages to the works council.

The appeal judgment reported that the company began consulting the works council on 25 September 2025, more than two months after the first court order.

What an auditable version would have shown

According to the judgment, Groupe Moniteur argued that employees had already been using ChatGPT themselves and that the company had simply introduced rules for its use. It also argued that DIGI brought together functions already available in other software and that employees were not required to use it.

The court rejected those arguments. It found that employees’ earlier use of AI did not remove the requirement to consult the works council, particularly as that use had previously been hidden or even prohibited. Making DIGI optional did not remove the requirement either.

In this entry’s reading, the company needed to provide the works council with written information about the AI tools before making them available.

That information should have identified which tools were available, who could use them, what tasks they supported, which roles they affected and how their use would be measured.

A dated record of the consultation would have shown what information the works council received, when it received it and whether the consultation had been completed before the tools were introduced.

The sources reviewed do not show what records the company held about the proposed rollout or what written information it gave the works council when the consultation began.

Where the gap was

The court reported that the AI tools could significantly affect working conditions, particularly for journalists and editorial secretaries. It found that the tools could change the nature and volume of their work. The court also held that when a project is introduced in stages, the works council must be consulted at each stage.

In this entry’s reading, the company had made the AI tools available before consulting employee representatives whose work the tools were likely to change. The works council took the matter to court, and the company began the consultation after a judge had ordered it to suspend the tools.

A ConductRecord keeps a record of each use of an AI tool, including which tool was used, what task it performed and what it produced.

A MetricRecord is designed to turn those records into signed aggregate figures. For example, it could show how often employees were using AI to transcribe interviews, edit articles or suggest headlines. The works council could use that information to examine how the tools were being used and which newsroom tasks might be affected.

ConductRecord and MetricRecord are Headlights designs.

What governance should have looked like

The court reported that DIGI and ChatGPT fell within the provisions on new technology under article L. 2312-8 of the French Labour Code. It held that Groupe Moniteur was required to consult its works council before introducing the tools, regardless of whether employees had used AI before or whether DIGI was optional.

The court found that the failure to consult was a manifestly unlawful disturbance and confirmed the order suspending the tools until the consultation was completed.

Where an employer introduces AI tools that could change how people work, best practice would be to consult the works council before making the tools available. The council should receive written information explaining what the tools do, which employees and tasks could be affected, and how the company intends to measure their use and impact.

The consultation should take place before employees are given access to the tools. If the rollout takes place in stages, the company should consult the works council at each stage.

A joint union statement published in December 2025 reported that the company had reopened the tools after the works councils issued unfavourable opinions on the project. The sources reviewed do not include the company’s account of that reopening.

Failure Pattern: an employer made an AI drafting assistant available to its newsroom and authorised the use of ChatGPT before consulting the works council that French law required it to consult, and a court suspended the tools until it did.

Governance Principle: an employer bringing AI into how work is done should be able to show that the people whose work it changes were consulted before the tools were made available, and on what written information.

The reference implementation of ConductRecord and MetricRecord is open source. It lives at github.com/saffronandindia/headlights-oss, Apache 2.0 licensed and free to install. The repository is public now.

Sources

The mailing list

Fresh incident reports every week. One email to match.

We add new incidents to the library regularly, and send a single short email each week with what's new. The library stays free and open; this is just how you keep up with it.

Sent through Substack. Unsubscribe in one click.

The record

An auditable system would have produced a signed, tamper-evident record the moment this happened: what the system did, the version that did it, the basis it acted on, and the action taken, and Groupe Moniteur could have produced it on demand.

This is the record the system as deployed did not produce in a signed, auditable form.

What this teaches
Capture what happened when it happens
What the system did, the version that did it, the basis it acted on, and the action taken, recorded at the moment, not reconstructed after.
Sign it, so no one has to trust the record-keeper
A tamper-evident entry. Edit it later and the signature breaks. The record does not ask for the benefit of the doubt.
Make it verifiable by anyone
A court, a regulator, a customer's lawyer can check the record themselves, without taking the company, or us, at our word.

Headlights summarises publicly reported AI incidents. All summaries are independently written, attributed to their original sources, and intended for research and educational purposes. Allegations are identified as such until established through official findings.

Last reviewed September 2026. This report is based on the sources listed above and reflects information available at the time of review; later developments may not be captured. Where a person is described as charged with or alleged to have done something, that allegation is unproven unless a conviction or a court or regulatory finding is stated. Headlights publishes journalism and commentary, not legal advice.

Want to write back?

Direct to my inbox.

ellie@useheadlights.com →