Hackathon Women in AI Safety - Highlights 👩‍🔬
Bridging the gap between 42 students and AI research !
This year, 42AI celebrated International Women’s Rights Day by hosting a jam site for the worldwide Women in AI Safety Research Hackathon, a global event organized by ApartResearch that gathered more than 350 participants worldwide !
For a little more than a weekend, contenders worked at crafting a research paper on one of the challenge’s proposed tracks:
- Mechanistic Interpretability
- Social Sciences
- Public Education
Participants from our 42AI Paris jam site have demonstrated great interest for these topics, which eventually led to 5 team submissions covering as diverse subjects as AI biases, disinformation and fake news, guardrails for medical AI chatbots, neuronal pathways coverage or behavioral nudges to raise awareness !
Without further adue, we are happy to shed light on these works :
- “AI bias in resume screening” by Inès Aliane and Mauro Abidal.
- “Medical Agent Controller” by Quentin Marquet, Ouafae Moudni, Shakthivel Murugavel, Xavier Charles and Elise Racine.
- “Mechanistic Interpretability Track: Neuronal Pathway Coverage” by Garance Colomer, Luc Chan, Sacha Lahlou and Leina Corporan Miath
- “Identification if AI-generated content” by Yvan Bondar & Bakyt Naurzalinov
- “AI-Powered Policymaking: Behavioral Nudges and Democratic Accountability” by Jana Dagher & Michel Chbeir
Congratulations to all participants, for most of whom this was a first attempt at writing a research paper ! đź‘Źđź‘Źđź‘Ź
This was a first hackathon for us involving research, but definitely not our last ! We’ll be happy to host this kind of competition again, and continue to fill the gap between Academia and our members !
The 42AI team 🚀
