Je sors de près de 3h de briefing et d’échanges avec le Président de la République et les services de renseignement et de sécurité. Du haut de mes 50 ans, je n’ai rarement ressenti tel vertige. Vertige de la gravité des situations au Moyen-Orient et en Ukraine, de la folie prédatrice qui mène le monde au chaos. Vertige des risques et conséquences sur notre sécurité, notre économie, notre pouvoir d’achat. Vertige de tous ces candidats qui rêvent, sans retenue, sûrs de leur fait et de leurs postures, de gouverner la France dans cette époque et cette situation historique et grave. Notre démocratie, notre pays doivent déjà faire face aujourd’hui et devront faire front demain. Plus que jamais je me battrai, à ma modeste place, pour que la France ne cède pas son destin aux egos délirants, aux va-t-en guerre de salon, aux pacifistes de pacotille et aux soumis de la facilité et de la peur. La France doit plus que jamais être forte. Nous devons tous y contribuer. L’avantage du vertige, c’est qu’il vous rappelle de ne pas tomber. @PartiRadicalG
AI has enormous potential, but a growing number of those building it are warning that development is racing ahead of our understanding of the risks.
We can’t afford to ignore their concerns.
We need genuine international cooperation & a coordinated global effort to make AI safe, transparent, accountable, with human dignity at the centre.
Last month I wrote about how we can build a positive and safe future for everyone: https://t.co/eoLGVY8yad
Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens.
The reality is:
- People won't want to use agents that are misaligned with them and that don't do what they ask, so labs have a strong natural incentive to make their models more aligned.
There is a lot of debate about slowing progress on capabilities until alignment catches up. My view is that trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models. Any lab that doesn't focus on alignment will fall behind.
- Labs face significant liability if their models cause harm, so they have a strong incentive to prevent this as well.
Meta delayed shipping Muse for several months to focus on safety and security. We didn't call for everyone else to do this before we would. We just did it as part of our day-to-day work because it was clearly the right thing for people and for us. I'm proud of the security foundations we've built.
- Engaging independent evaluators and advisors is industry best practice. MSL already does this today in several areas because it helps produce better work. Other labs can just do this too. In general, it would be helpful for there to be a larger and more diverse ecosystem of evaluators.
- Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely. Meta has made this commitment and other labs can do this as well.
I believe the key to building a positive future for everyone is maintaining the right balance of power. This is within our power to do.