Skip to content

Commit 9c2c11e

Browse files
committed
MV - favicon
1 parent ff27528 commit 9c2c11e

11 files changed

+117
-1
lines changed
Lines changed: 54 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,54 @@
1+
---
2+
# ************************************************************************
3+
# FrontMatter
4+
# How to create your post, see : _posts/README.md
5+
# ************************************************************************
6+
title: Fresque de la SΓ©curitΓ© de l'IA x CeSIA
7+
date: 2025-25-03 18:00:00 +/-TTTT
8+
categories: [Workshop,AI_Safety]
9+
tags: [42AI,Γ‰vΓ©nements,AI_Safety]
10+
description: "DΓ©couverte des enjeux de l'AI Safety"
11+
image:
12+
path: /assets/img/events/fresque_ia_safety.png
13+
date: 2025-03-18 13:37:00 +/-TTTT
14+
# Author : will default to 42AI if not filled.
15+
author:
16+
17+
---
18+
19+
# 42AI x CeSIA - Fresque de la SΓ©curitΓ© de l'IA πŸ”¬
20+
21+
### πŸš€ 42AI accueille la Fresque de la SΓ©curitΓ© de l’IA en partenariat avec le CeSIA ! πŸ€–
22+
23+
​Le [CeSIA](https://securite-ia.fr) est une **organisation dΓ©diΓ©e Γ  l'Γ©ducation, la recherche, et la diffusion d'informations sur les enjeux de l’intelligence artificielle**.
24+
25+
### Comprenez-vous vraiment les enjeux de sΓ©curitΓ©, de fiabilitΓ© et d’éthique de l'Intelligence Artificielle (IA) ?
26+
27+
​Participez Γ  notre atelier collaboratif et ludique pour explorer et comprendre les enjeux et dΓ©fis pour se les approprier.
28+
29+
​Dans un contexte d’augmentation toujours plus importante des capacitΓ©s des IA, des risques humains et sociΓ©taux apparaissent ou sont amplifiΓ©s.
30+
31+
​Sans prΓ©requis techniques, cet atelier ludique et pΓ©dagogique vous permettra de :
32+
33+
​- Construire collectivement une fresque qui rΓ©vΓ¨le les liens entre les capacitΓ©s de l'IA, leurs implications et les risques associΓ©s,
34+
35+
- ​Participer Γ  des Γ©changes enrichissants en petits groupes, accompagnΓ©s par nos facilitateurΒ·ices,
36+
37+
- ​DΓ©velopper une comprΓ©hension claire des enjeux de sΓ©curitΓ© de l'IA.
38+
39+
​Notre atelier vous guide Γ  travers les trois types de risques principaux de la sΓ©curitΓ© de l'IA :
40+
41+
​- les impacts systΓ©miques sur notre sociΓ©tΓ©,
42+
43+
- ​les dΓ©fis du dΓ©salignement entre les objectifs humains et ceux des IA,
44+
45+
- ​les risques d'utilisations malveillantes.
46+
47+
​Comprendre les enjeux de sΓ©curitΓ© devient essentiel pour tout citoyen Γ©clairΓ©.
48+
49+
​Notre atelier vous donne les clΓ©s pour : dΓ©crypter les implications de l'IA dans notre sociΓ©tΓ©, contribuer aux dΓ©bats actuels de maniΓ¨re informΓ©e et Γ©changer avec d'autres participants dans un cadre bienveillant.
50+
51+
HΓ’te de vous y voir !
52+
53+
BESS - CafΓ© des Assos
54+
Mardi 25 Mars - de 18h Γ  20h
Lines changed: 63 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,63 @@
1+
---
2+
# ************************************************************************
3+
# FrontMatter
4+
# How to create your post, see : _posts/README.md
5+
# ************************************************************************
6+
# DO NOT EDIT THE FOLLOWING LINES
7+
layout: post
8+
categories: [Hackathon,AI_Safety]
9+
is_blog_post: True
10+
# ************************************************************************
11+
# EDIT THE FOLLOWING LINES
12+
13+
# // REQUIRED date : 2024-06-13 13:37:00 +/-TTTT
14+
# Only modify the date and hours elements, leave the +/-TTTT unchanged.
15+
date: 2025-03-27 13:37:00 +/-TTTT
16+
17+
# Author : will default to 42AI if not filled.
18+
author:
19+
20+
# Tags: add a 3rd TAG to the following list to evoke your article's content thematic.
21+
# If an existing tag works with your article, please use it in priority.
22+
# Existing : Generative_AI ; AI_Safety ; AI_Entrepreneur ;
23+
tags: ['Hackathon', '42AI', ]
24+
25+
# // REQUIRED Description : text that will display under the title // + SEO // Keep it short and relevant
26+
description: Discover our participants' works!
27+
# Image: the illustration image that will accompany your article on links and in the article itself.
28+
# Images should be of dimensions ratio 3:1
29+
# Images can point to external urls
30+
# If you don't have an image, leave this item unchanged.
31+
# You can place your custom image in the `/assets/img/blog/` folder. If you do, don't forget to edit the following
32+
image:
33+
path: /assets/img/events/hack_womenInAiSafety_retex.png
34+
35+
# // REQUIRED Title : Keep it short, catchy and consistent with your article
36+
title: Hackathon Women in AI Safety - Highlights πŸ‘©β€πŸ”¬
37+
---
38+
39+
## Bridging the gap between 42 students and AI research !
40+
41+
This year, 42AI celebrated **International Women's Rights Day** by hosting a jam site for the worldwide
42+
[Women in AI Safety Research Hackathon](link), a global event organized by [ApartResearch](https://apartresearch.com/sprints/women-in-ai-safety-hackathon-2025-03-07-to-2025-03-14) that gathered more than 350 participants worldwide !
43+
44+
For a little more than a weekend, contenders worked at crafting a research paper on one of the **challenge's proposed tracks**:
45+
- **Mechanistic Interpretability**
46+
- **Social Sciences**
47+
- **Public Education**
48+
49+
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 !**
50+
51+
Without further adue, we are happy to shed light on these works :
52+
53+
- **["AI bias in resume screening" by Inès Aliane and Mauro Abidal](https://apartresearch.com/project/ai-bias-in-resume-screening).**
54+
- **["Medical Agent Controller" by Quentin Marquet, Ouafae Moudni, Shakthivel Murugavel, Xavier Charles and Elise Racine](https://apartresearch.com/project/medical-agent-controller).**
55+
- **["Mechanistic Interpretability Track: Neuronal Pathway Coverage" by Garance Colomer, Luc Chan, Sacha Lahlou and Leina Corporan Miath](https://apartresearch.com/project/mechanistic-interpretability-track-neuronal-pathway-coverage)**
56+
- **["Identification if AI-generated content" by Yvan Bondar & Bakyt Naurzalinov](https://apartresearch.com/project/identification-if-ai-generated-content)**
57+
- **["AI-Powered Policymaking: Behavioral Nudges and Democratic Accountability" by Jana Dagher & Michel Chbeir](https://apartresearch.com/project/ai-powered-policymaking-behavioral-nudges-and-democratic-accountability)**
58+
59+
Congratulations to all participants, for most of whom this was a first attempt at writing a research paper ! πŸ‘πŸ‘πŸ‘
60+
61+
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 !
62+
63+
The [42AI](https://42-ai.github.io) team πŸš€
964 KB
Loading
221 KB
Loading
-10.9 KB
Binary file not shown.
-28.7 KB
Binary file not shown.
-9.91 KB
Binary file not shown.
-735 Bytes
Binary file not shown.
-1.63 KB
Binary file not shown.
-15 KB
Binary file not shown.

0 commit comments

Comments
Β (0)