Launching a French AI safety community through university outreach and local events to connect students, researchers, practitioners, and policymakers.
Launching a French AI safety community through university outreach and local events to connect students, researchers, practitioners, and policymakers.
Arthur is an engineering student who wishes to raise awareness on AI safety risks to the youth, hoping the thinkers of tomorrow would find an interest in tackling the challenges rising along AI innovation. Julia is a finance and strategy student, currently conducting research and engaged in making AI safety accessible to youth through game-based learning, her aim is to contribute to solving the present and future AI’s risks. Jeanne is an international relations student, whose focus on political violence led to research AI governance and safety, her objective is to support talents finding resources and a network of collaborators to accelerate into the field. Together, we form a bilingual technical-governance tandem and hope to share our knowledge with peers to contribute a French field-building effort into the broad AI safety community.
This pilot project aims at creating a feedback loop to draw more attention to AI safety problems, accelerating talents into the broader ecosystem and building a long-term French community. To this end, we will take several strategic steps, starting in August for our pilot launch, which is estimated to last 4 months:
First, we will contact different universities and student associations to organise talks and workshops with tailor-made contents based on the audience. Each member has a different portfolio, with theoretical and practical orientation, which serves the purpose of creating a varied but targeted portfolio. Julia has a political science and economy background, as well as her connections to Sciences po Paris, Pause AI, GPAI Policy Lab and CharabIA. Jeanne has an international governance perspective and insights on China, and connections with Effective Altruism and the LessWrong community. Arthur will handle the technical presentations for engineering and computer science students. We plan to develop educational materials for business, philosophy and communication students, with the help of future collaborators, who we hope to integrate in the team along the way.
Then, a priority for us is to hire a communications’ officer, ideally under an internship contract. They would support the creation of a visual identity, as well as initiating a strong social media presence to have a direct contact with youth online and attract new members. The benefits of having a full-time intern rather than a part-time volunteer are to have an engaged and committed team member who would follow clear objectives, while helping a student who is interested in AI to professionalise in the field, be independent and create a portfolio from an idea. It is a simultaneous support of accelerating talents into the AI ecosystem, as it often needs agentic communication-oriented people who are able to create from day one. French internships that last over 2 months are supposed to be compensated by a minimal stipend of 4.5 euros per hour.
Third, starting in September, we would like to start registering Apart Research hackathon hubs, as it has never been done in France so far. They will allow the gathering of interested researchers and spark collaborations in the field of AI safety. This will start the trigger of the feedback loop, as people who received talks could start committing time and energy to AI safety. While Apart Research provides feedback, which helps in improving and iterating research projects and prototypes, the registration of hubs under Horizon AI will also help accelerate talents into the field. They could, in turn, be recruited into the association and start raising awareness among future decision-makers about AI risks and dangers.
Finally, to mark a milestone in our launch, we aim to organise an AI Safety France event in November. This would close the loop by connecting together people from the ecosystem, from researchers to entrepreneurs and policymakers. The French-speaking AI safety ecosystem is fragmented and needs such initiatives to show future generations different opportunities and at the same time bridging the gaps between practitioners who need to be in each other’s direct network to prevent or address AI safety challenges, such as X risks.
The French AI safety ecosystem is highly fragmented despite the quality of leading AI startups, research universities and public institutions. Students, early-career professionals or people who wish to transition often do not consider AI safety, or struggle to find opportunities to access resources or collaborations to take their first steps into the field. We believe that researching and addressing socio-technical challenges to reduce catastrophic AI risks is a coordination challenge that requires strong links between different domains.
The goal of our pilot is to weave the web within this ecosystem by fostering the impulsion of a feedback loop to increase both the number of talents and diversity of fields able to contribute to AI safety. We aim at touching our peers first, through university talks, local hackathon hubs, then starting larger scale community events. The ideal outcome for us is to lower barriers to entry, build an acceleration pipeline and foster interdisciplinarity (technical to philosophical talents are needed in the field). Increasing connectivity in the French-speaking side with the international AI safety community is the ultimate goal to build a stable and long-term organisation where AI development and governance are one instagram comment away rather than a few months of unmentored blind exploration. Indeed, according to Pause AI, there is a lack of cohesive social engagement in the field, which we would attempt to fill ().
Team Member
https://docs.google.com/spreadsheets/d/1p0IsR-QzHo8t5kVl8S43FxVXi8LSj11gjngEug0nv5I/edit?usp=sharing
The conservative estimates allow for a fair kick-starting of the project, but if more money is collected, priority will be given to university outreach and communication, as they are the basis of our field-building project.
Alone, our pilot does not directly reduce AI x-risks, but supporting the next generation of thinkers could further the cause by evoking new passions in high-impact careers such as policymaking. Field building is a crucial step, and a current bottleneck in AI safety, notably in Europe. It could also participate in improving global coordination and effort to reduce existential AI risks.
This seems like a great project! All the very best :)
Thanks !