Skip to content
Submit incident
AII-1002-63131
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.

French AI chatbot Lucie shut down after basic errors

French-language AI chatbot Lucie, developed with government support, was shut down approximately three days after launch on 23 January 2025 due to providing obviously incorrect answers to basic questions, including mathematical errors and nonsensical biological claims. The developer acknowledged the release was premature and the system lacked adequate testing.

23 January 2025 Β· Linagora Group Β· Reliability and hallucination
23 January 2025
AII-1002-631312 sources ↓
Quick facts
Category
Reliability and hallucination
Country
France
Source
Liberty Times Net
Date
23 January 2025
Drafted by
Raidu Incident Detection

What happened

Lucie, a French-language AI chatbot developed by Linagora Group with government support, launched on 23 January 2025. Within days, users reported the system providing clearly incorrect answers to simple questions.

When asked about 'cow's eggs', Lucie responded that they are 'also called chicken eggs' and are edible eggs produced by cows containing protein and nutrients. When asked to calculate 5Γ—(3+2), Lucie answered 17 instead of 25, failing to apply the order of operations.

Reported impact

Users shared multiple examples of the system's incorrect responses on social media, generating criticism and public disappointment with the system's performance.

Index analysis

The system failed on elementary arithmetic and semantic reasoning tasks, confusing biological categories and mathematical order of operations. The failures suggest inadequate testing before deployment.

Governance controls

Linagora Group acknowledged the model was in an early research stage and the release was premature. The company stated it should have more clearly communicated the model's limitations to users. The system was shut down after approximately three days in operation for further testing and updates.

Classification

Category
Reliability and hallucination
Industry
Government, security and defence
Country
France
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Reputational
Harmed
General public
Failure modes
Hallucination, Agent misaction
Consequences
System suspension or termination
AI systems
Lucie

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Liberty Times Net
ζ³•εœ‹AIθŠε€©ζ©Ÿε™¨δΊΊLucie η΅¦ε‡Ίι›’θ­œε›žη­”θ’«δΈ‹ζžΆ - εœ‹ιš› - θ‡ͺ由時報電子報
28 January 2025
news.mydrivers.com
French AI chatbot Lucie shut down after basic errors