Sasha Baker’s Post

This week, OpenAI joined other leading AI companies and representatives from the international network of AI safety institutes in San Francisco at a convening about the safe development and deployment of this revolutionary technology.  Around the world, ChatGPT and other generative AI tools are creating new economic opportunities and helping people solve hard problems in areas like healthcare and education. But to fully maximize AI’s benefits, the U.S. and like-minded democracies need to develop a shared set of interoperable standards that prioritize both safety and innovation. Getting the balance right will require harmonizing technical standards globally to avoid creating a patchwork of competing safety evaluations and regulations that could stifle innovation, prevent AI-related economic gains, and limit access to the technology. Similarly, we believe that U.S.-based companies should be able to conduct pre-deployment testing at home, and that there should be mutual recognition across jurisdictions. If a U.S. company meets safety requirements here, that should be enough to satisfy other countries with safety institutes of their own, and vice versa. Reciprocity will streamline the evaluation process while fostering trust among international partners. That sort of cooperation will be particularly important when it comes to coordinating research and testing of advanced AI models in critical areas of national security, including cybersecurity and mitigating against chemical, biological, nuclear, and radiological (CBRN) risks, which are top priorities for OpenAI. We want to thank Commerce Secretary Gina Raimondo and Secretary of State Antony Blinken for organizing the summit, as well as the participants who came to San Francisco from around the world. We’re committed to the safe and responsible use of AI and look forward to working with our international partners at the AI Action Summit in Paris next February. #AISafety

Elizabeth Kelly

Director, U.S. AI Safety Institute | Policy Leader | Startup Executive

1w

Grateful for your support and leadership in advancing the science of AI safety.

Malik Ahmad Jalal

Founder Cordoba Ventures & Zambeel Partners I Executive Director Goldman Sachs I Harvard I L.S.E I I.C.A.E.W I Venture Capital I Healthcare I Logistics

1w

Sasha - congrats. How do you end up at some of the most interesting assignments in the most interesting times!

Like
Reply
Samantha Neakrase

Senior Director @ NTI | Nuclear Policy Expert

3d

I think many of us feel a lot better knowing you are working on this!

Like
Reply
See more comments

To view or add a comment, sign in

Explore topics