OpenAI and AI Futures

I recently had the pleasure of attending an OpenAI event, where a number of academics, government types, and entrepreneurs convened to discuss artificial intelligence. The questions ranged from the changing nature of work, to how AI can play into international relations, to transformation of industries, to what the state-of-the-art is.

Obviously, artificial intelligence is a complex and extremely timely issue, one that inspires a wide range of opinions in people.

In fact, I often describe people's opinions about AI as functioning like either a mirror or a Rorschach Test. In the case of a mirror, a speaker looks at the potential of AI as a reflection of themselves. What would I do if  I had an AI? In the case of a Rorschach Test, a speaker looks at the potential of AI via how they think other people in the world work. What would they do if they had an AI?

Reflections range from Star Trek universes — where robots produce all material needs for all human life — to the kind of dystopias Hollywood more commonly produces. It’s an odd part of life that dystopias and fears sell much better than hopes; I myself tend to be quite optimistic about the future we can create.

And this is part of why I’ve been happy to support OpenAI since its founding. It contributes a critical element to the overall development of AI.

For decades, universities sustained AI research.  Now, simply amazing progress is coming from a set of key company labs — everything from Alpha Zero to image recognition to language processing and more. Most of the researchers who work at these companies are great people — not just smart, but also appropriately concerned about the results of their work on society. Some of these folks I count as close friends, and their contributions to AI will continue to be crucial to the progress we as a society make in this realm.

However, as great as these corporate labs and their leaders are, it’s essential to have counterbalancing institutions as well — like OpenAI, which pairs the inherent liberty of academia with the resources enjoyed by corporate players. All organizations are inevitably beholden to their structure. Corporations, for example, focus on a revenue and profit line, a competitive position, and a strategic plan for growing their economic capabilities.

Like corporations, OpenAI is at the forefront of AI research and has access to massive compute scale. But unlike corporations, OpenAI is a non-profit. Its mission puts humanity first — rather than business goals. And its structure allows it to attract philanthropic support while also providing public accountability.

Because OpenAI is independent from companies and their competition, it can focus on the best outcomes for humanity and society. For example, OpenAI can provide a set of tools for all companies and all countries. It can serve as an independent convener when companies in the industry need to discuss standards or safety issues. It can invest as much in learning, safety, and collaboration as it feels will help the development of great AI — with no worries about revenue over cost models.

In short, OpenAI focuses on AI for everyone, and for the world.

Of course, it isn’t just the non-profit structure and the mission that makes OpenAI so valuable. It is essential that it has attracted a set of great people who have decided that this is their mission.

When I went to the OpenAI event, I already knew that the people who work there are great technologists on a mission. I already knew that they're focused on how to shape the technology to human benefit. What I learned is that they're also committed to getting many other voices into the conversation, to make these discussions as diverse and inclusive as possible. Attendees included the new minister of AI for Dubai (who is a friend), to academics, to intellectually dedicated investors.

OpenAI is not about open source development — the technology needs to be developed and shaped for the right outcomes and safety. But it is about open benefit to humanity. And like all 501c3s, OpenAI is effectively owned by the public interest. In the same way that the Internet's creators provided a set of open standards and protocols that lots of people could use to develop upon it, OpenAI is giving the world open tools related to AI.

I fund OpenAI because it’s the key research and development organization outside of the dominant commercial platforms. It’s similar to why I spent over a decade on the Mozilla board.

A structure that blends public interest and philanthropy. A mission focused on AI for the entire human ecosystem. Excellent technologists. The unique technical base for general AI tools that are not beholden to any one major commercial platform. A commitment to be in conversation with the world. And that's why I'm increasing my support for OpenAI, because its unique capabilities can also move the entire AI field forwards, both directly and indirectly.

As technologists devoted to improving the world, we want to anticipate trends before they happen. Then we want to help shape them for the best impact on humanity. Sometimes that consists of briefing the public and policymakers — some of which happened at this OpenAI event. Sometimes that consists of creating a key technology safety protocol. OpenAI has the ability to help AI move forward in ways that ultimately benefit all humanity. Like another non-profit I support, the Ethics and Governance of Artificial Intelligence Fund, it functions as a strong role model for how technologists and the tech industry should engage as they pioneer the future of society.