OpenAI’s CEO of Applications Fidji Simo on Her Vision for Growth

By Harry Booth

In May, Sam Altman announced that tech executive Fidji Simo would join OpenAI under a newly-formed title: CEO of Applications. “Fidji will focus on enabling our ‘traditional’ company functions to scale as we enter a next phase of growth," Altman wrote. Despite rapidly becoming one of history’s most valuable startups, OpenAI runs significant losses, with the organization estimating it lost $5 billion last year. Simo faces the daunting task of steering the company toward profitability.

She brings a resume made to meet the moment. The former CEO of Instacart, an online grocery platform in which Altman was an early investor, Simo wrangled the company out of a post-pandemic slump into a period of growth. Before that, she spent a decade at Facebook, where she was instrumental in building its monetization engine, launching the news feed ads and video products that transformed the social network into a financial juggernaut. 

OpenAI is, in many ways, very different. Despite recent efforts to scrap its unique corporate structure, the company remains bound to a non-profit board—on which Simo previously sat. That board is tasked with ensuring the business acts in accordance with its mission of ensuring artificial general intelligence—or AI systems whose capabilities surpass those of humans—benefits all of humanity. “If you look at social media, their goals were always around engagement [or] time spent, whereas for us, we do not optimize for engagement. We optimize for the ability to be useful to people,” Simo told TIME in July. “My core belief is that any technology, but AI in particular, should empower people, and that this empowerment should be distributed to all, instead of concentrated for a few."

This interview has been condensed and edited for clarity.

In a recent OpenAI blog post titled “AI as the greatest source of empowerment for all,” you wrote that “AI can give everyone more power than ever,” but also that these opportunities “won't magically appear.” Given your experience leading an AI transformation at Instacart and your year on the OpenAI board, what are the opportunities you've spotted to unlock that power?

I really define myself as a pragmatic technologist. I grew up in a place that wasn't steeped in technology—the South of France—and really, when I discovered technology, what fascinated me was how we can apply it very concretely to solve problems in people's lives. And when I look at AI, it's such a versatile technology that I believe it's going to be able to solve a lot of different problems.

The reason I focused on empowerment is because if you look historically, a lot of technologies have had the potential to put power in the hands of people, and some certainly have. But sometimes you run the risk of reinforcing power in the hands of people who already have it—who already have more money, more connections, more time, all of the sources of leverage. Whereas when I look at AI, I believe that if we do our job right, we're really going to be able to put this empowerment and these opportunities in the hands of everyone.

I mentioned a couple of cases in my essay. Some of the ones I'm really passionate about include health in particular. When I look at how ChatGPT is being used, I really see the potential for having people regain their power when it comes to their health and their bodies. I found it astounding that nine out of 10 people do not understand health information well enough to make informed decisions on the most important part of their lives. So when you walk into a doctor's office, the power imbalance between a doctor that has the knowledge and someone on the other side who needs to make decisions for the most important part of their life, and doesn't have the right knowledge, is really striking. I believe that with tools like ChatGPT, we can rebalance that power.

Same thing when it comes to knowledge and education. I believe that we are going to end up creating a system where everybody can learn, and that is one of the biggest ways in which you can get access to more economic opportunities in the future.

With your role as CEO of Applications, does the vision lean towards OpenAI becoming a direct-to-consumer app builder, or more towards enabling a broad ecosystem of others to build on your models?

I think it's both. When I think about my role, it's really turning breakthrough research from our incredible research team into products and services that people are going to be able to use in the world. Whether that's done by us or with us by building a platform that others can build upon will really depend on the use case. 

There are some things, like ChatGPT, where having a first-party app that can really show the way AI can be distributed to people is incredibly important, but we are not going to be able to build everything. All the use cases I mentioned are actually use cases that are ripe for other startups to build upon. And so we are really going to do both: develop a first-party app so that we can really show the world how this research can best land in the hands of people in a way that empowers them, but also give a lot of companies the tools to build the best applications on top of that.

Will this newer side of the business operate independently under your leadership?

I think the magic of OpenAI is the tie between these things: the tie between the research lab and the way to deliver these things into the world through applications, the tie between our compute team and how we deliver things in the world. So I am not planning on taking my org as a silo. 

I am responsible for that part of the business and able to make decisions with autonomy, so that Sam can really focus on the research and compute aspects of the business, which is what he's always wanted to focus on.

Your past employer, Facebook, faced intense criticism for how its business incentives—including on some of the platforms you helped shape like video and ads—fueled user addiction, divisive content, and even violence. As you step into this new role, how will you proactively prevent similar pitfalls at OpenAI? What lessons from Facebook are you bringing to ensure that AI's business incentives don't undermine its core mission of “benefiting humanity?”

Two things: One, we have very different goals, and I think goals determine a lot of actions at the company. If you look at social media, their goals were always around engagement [and] time spent, whereas for us, we do not optimize for engagement. We optimize for the ability to be useful to people. In fact, we believe that if we do our job right, tools like ChatGPT will be able to do so much more on your behalf, autonomously, and add value to your life without you having to be in the tool all the time. Maybe people are actually going to interact with the tool less in the future, but get much deeper usefulness from it. 

The second thing is that I think in the early days of social media, people, including myself, were all looking towards the positive, optimistic use cases of the technology, and we were always on our back foot when it came to the negative use cases. With AI, and in particular at OpenAI, we have taken an approach of being exceptionally proactive in identifying sources of potential future harms and, in fact, have been very vocal since the beginning about the risks of this technology. When you look at what we build, we have as robust of a roadmap on safety tools as we have a roadmap on new research breakthroughs. And that mentality of really thinking of safety as a very, very big part of what we do—which is as important as developing the capabilities themselves—is something that is completely different here than at any other company I have been exposed to.

So, the combination of these two things—the right objective, combined with a lot of focus on safety and actually putting resources behind it—is really what I think can help us fare better. 

At the same time, I want to acknowledge that this is a completely new technology, and there will be things that surprise us. That's why we have a philosophy of iterative deployment, so that we can take this technology, put it in front of people, learn very quickly how they use it in ways that we may not have anticipated, and then react incredibly quickly by taking these things very seriously and putting the right guardrails in place, so that we continue on our trajectory of being incredibly helpful to people and minimizing all of the downsides of the technology.

Your background in targeted advertising at Facebook and Instacart involves nudging human behavior to drive sales. In your manifesto, you talk about nudging human behavior, albeit in a totally different direction towards things like healthy habits. This raises questions around human agency. Why should people place that trust in OpenAI? How do you make sure that you're genuinely empowering individuals, rather than subtly influencing or controlling them?

I think knowledge is power, and I think the line that you don't want to cross is the line between giving people knowledge versus being paternalistic and telling them what to do. I saw that throughout my career, even at Instacart. We obviously are at the center of the relationship that people have with food, and I've always taken the approach of wanting to tell them on every ingredient that they buy what it is that they're eating, being very transparent on what they're about to put in their body, but never telling them, "Hey, you really shouldn't be eating the Nutella," because I love my Nutella, and I would be very upset if someone told me not to eat it. 

I think that's the line that we're working with with ChatGPT, which is really helping people with their goals and, over time, being more explicit in really understanding their goals and then helping them with knowledge and information to reach their goals, but never imposing our own goals. That is not our role. Our role is to really empower people to realize theirs.

Sam Altman has said your appointment signals OpenAI's entry into its next phase of growth. When you think about the coming years, what excites you most?

I would say the big transformation that we're going through right now in terms of products we put out into the world is this transformation from chatbots to true agents. We just launched ChatGPT Agent. It is pretty magical. I was just able to book a weekend trip, which would have taken me many hours, and here it just took really a few minutes. When I started playing with that, I realized there is a fundamental shift that is happening from using ChatGPT as a tool that mostly gives you information to a tool that actually does things on your behalf. We're early in the development of that technology, but I think it really gives us a glimpse of what the future is going to look like, both on the consumer side and the enterprise side. 

[With] a super-assistant that's acting on your behalf, that has your back, allows you to reach your goals, and can do that across all parts of your life, I think we're just starting to see a really, really incredible future of freeing people up from doing a lot of things they don't want to do, [to] allowing them to get access to knowledge they wouldn't get access to otherwise, [to] having more economic opportunity, [to] recovering their time. And all of that is based on these technology breakthroughs that we're witnessing.

Disclosure: OpenAI and TIME have a licensing and technology agreement that allows OpenAI to access TIME's archives. 

(To receive weekly emails of conversations with the world’s top CEOs and decisionmakers, click here.)

Must-read stories from TIME Business:

• A Court Has Ruled Most of Trump’s Tariffs Are Illegal. Here’s What’s Next

• The AI Summit Where Everyone Agreed on Bad News

• How Trump’s Punishing Tariffs on India Could Impact American Consumers