STEVE INSKEEP, HOST:
We have a skeptical look at one of the most famous companies of our time. Karen Hao wrote it. She's a journalist who first covered OpenAI six years ago when it was promising to develop artificial intelligence ethically.
KAREN HAO: It was founded as a research nonprofit meant to do this kind of fundamental research for, as it said, the benefit of all of humanity. And it was meant to open source that research to be for the public benefit.
INSKEEP: The word open was right there in the name. OpenAI had the backing of tech giants, including Elon Musk, who worried about AI's future effects on the world. So Karen Hao arranged to spend time in the company's offices as a young reporter for the MIT Technology Review. Her experience led her to ask questions which eventually led to a book called "Empire Of AI." It explores the company's research, its development of ChatGPT, its firing and restoration of CEO Sam Altman and more.
HAO: They were incredibly secretive, which didn't seem quite right because, first of all, that's the opposite of transparent. But second of all, why would you be secretive if you're a nonprofit that's purely doing research in the interest of the public? You would be secretive if you might have some kind of commercial intent. Also, they were incredibly competitive. When I was interviewing executives, they really emphasized that they needed to be No. 1 in AI progress. They needed to be first to what they called artificial general intelligence. And to me, there was an inherent tension in wanting to be first and wanting to be open.
INSKEEP: I feel that you help me understand that this is more than a clash of personalities. You have people who are worried about the future of humanity, the future of the planet, and thinking, we need to develop artificial intelligence in an ethical way, and that's why this should be a nonprofit. But we also need to be ahead of everybody else in order to shape things, and therefore we need to be really good, and therefore we need to get immense investment, and therefore we need to attach for-profit motives to the nonprofit. I mean, they kind of reason themselves out of their own original idea.
HAO: Yeah, that's exactly right. I think a lot of OpenAI and AI development in general is very much an ideological clash, not just personality clashes. Certainly, there are huge personalities, and they're all very egotistical, and they all want to be the one that makes AI in their image. But there are these deep-seated, kind of quasi-religious beliefs that also underpin the whole thing where some people believe that AI could have the potential to bring the world to utopia, and they want to run as fast as possible to that objective. And other people believe that AI has the possibility to destroy humanity, and they want to run fast to the finish line of building this powerful technology so that they're the ones that ensure that it's not another, quote-unquote, "bad actor" that is going to do it first and therefore lead to the demise of humanity.
INSKEEP: Who is Sam Altman and what did he bring to the table, given that he himself was not an inventor or a brilliant engineer?
HAO: He's a once-in-a-generation fundraising talent. That is his particular skill. And he's also a once-in-a-generation storytelling talent, which is effectively why he's so good at fundraising. He is able to paint these extremely persuasive visions of the future. He was already prominent within Silicon Valley, and Silicon Valley very much runs on stories and telling stories about the future. And one of the things that I sort of concluded through the reporting of my book is that when he says something to someone, what he's saying is more tightly correlated with what he thinks they need to hear than what he actually believes or the ground truth of the thing. He's able to say the things that really provoke people to kind of rally towards a general, broad, sweeping mission that he paints.
INSKEEP: So a few bottom-line questions. Has OpenAI remained the leader in artificial intelligence?
HAO: It's an interesting question because it depends on how you would measure it. In terms of name recognition, you could argue that they are the lead because most people have heard of ChatGPT. I recently was talking with some application developers who use AI models for their platform, for their service. And application developers are now moving towards a model where they are trying to make their platforms AI vendor-agnostic. And so in terms of the actual capability of the technology, it seems like increasingly AI has become commoditized. And OpenAI doesn't necessarily have a meaningful lead anymore.
INSKEEP: Can you tell me one more thing? You call the book "Empire Of AI," and you make an extended comparison to old-time colonial empires of the 1800s, say. What are you talking about?
HAO: We are essentially seeing the recreation of how empires of old used to work. Like empires of old, there were a small group of people at the top that were able to make decisions for everyone else around the world. And basically everyone else around the world didn't have agency, didn't have say. And they lived in the thrash of the decisions that were happening at the top, based on the whims of the people at the top. And we are now in basically the same situation.
I mean, the empires of AI, as I mention in the book, they're not as overtly violent as empires of old because we've had 150 years of moral and social progress. So they're not going to be using the kind of same violent methods to do the exploitation and extraction. But they are still exploiting lots of labor around the world, in the sense that they are starting to apply real pressure to people's ability to receive economic opportunity by creating technologies that automate away a lot of work.
And they're also seizing up a lot of resources to continue fortifying their empire. They're seizing up land for their data centers. They're seizing up data that people put online without the realization that it would be commoditized and used to turn a profit. And we have to be extremely cognizant because if we follow this path to its logical conclusion, democracy cannot survive in a world where the vast majority of people no longer have agency and say and control over their own lives.
INSKEEP: Karen Hao is a journalist and the author of "Empire Of AI." Thanks so much.
HAO: Thanks so much.
INSKEEP: We will note Sam Altman did not give Karen Hao an interview for the book. NPR reached out to OpenAI for comment, and they did not respond.
(SOUNDBITE OF DANNY PAUL GRODY'S "ON LEAVING") Transcript provided by NPR, Copyright NPR.
NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.
