Looking Back: Solenne Gaucher on Unbiasing AI

On Thursday, March 6, 2025, IHES welcomed around sixty attendees for a talk by Solenne Gaucher, Assistant Professor at École Polytechnique, on the topic of gender bias in artificial intelligence, organized on the occasion of this year’s International Women’s Day.

Among the participants were more than 30 high school students from Lycée Rosa Parks in Montgeron, accompanied by their teachers Mathilde Boucher, Clément Lhote, and Jean-Baptiste Rolez. Their visit was made possible through a collaboration between IHES and the Académie de Versailles, via Inspectors Xavier Gabilly and Catherine Gufflet. The experience left a lasting impression on the students, as shared by Mathilde Boucher:

“The students were so excited. They talked about their visit during the whole bus ride back, and their eyes were still full of stars when we made it back to Montgeron.”

The conference was also streamed online and followed remotely by around twenty additional participants.

In her talk, Solenne Gaucher began by defining artificial intelligence and the concept of algorithms, before introducing the three key goals of machine learning algorithms: efficiency, accuracy, and objectivity.

She then warned that today’s AI systems still fall short of these ideals and can lead to biased or even harmful outcomes. Through concrete examples, she demonstrated how gender bias can emerge in the suggestions and decisions made by machine learning tools.

After raising awareness of these challenges, she explained how machine learning algorithms actually work and explored where these biases originate. She concluded with an inspiring look at the crucial role mathematicians can play in designing fairer, more equitable algorithms.

The event wrapped up with a lively Q&A session, which continued informally during the post-talk reception.