TheChek News Flow

Open AI’s Altman and Mira Murati say artificial intelligence will be ‘the best tool humanity has yet created,’ but the key will be to manage the transition

In the fast-moving world of artificial intelligence, talk about the future tends to focus on something referred to as “artificial general intelligence”—the stage at which AI will be capable of doing any job a human can do, only better.

Perhaps no company is in a better position to talk about the potential benefits and risks of this so-called AGI than OpenAI (49%-owned by Microsoft MSFT 0.81%increase; green up pointing triangle), the creators of the popular chatbot ChatGPT. At The Wall Street Journal’s annual Tech Live conference last week, the Journal’s senior personal-technology columnist Joanna Stern spoke with OpenAI’s Chief Executive Sam Altman and Chief Technology Officer Mira Murati. Edited excerpts of their conversation follow.

SAM ALTMAN: The two things that will matter most over the next decade or few decades to improving the human condition are abundant and inexpensive intelligence: the more powerful, the more general, the smarter the better. I think that is AGI. And then, abundant and cheap energy. If we can get these two things done in the world, then it’s almost difficult to imagine how much else we could do. We’re big believers that you give people better tools, and they do things that astonish you. And I think AGI will be the best tool humanity has yet created.

With it, we will be able to solve all sorts of problems. We’ll be able to express ourselves in new, creative ways. We’ll make just incredible things for each other, for ourselves, for the world, for this unfolding human story. And it’s new, and anything new comes with change, and change is not always all easy.

WSJ: When will it be here, and how will we know it’s here?

MIRA MURATI: Probably in the next decade, but it’s a bit tricky. Often we talk about intelligence, and how intelligent this is, or whether it’s conscious, and sentient, and all of these terms—and they’re not quite right. Because they sort of define our own intelligence, and we’re building something slightly different.

WSJ: Some people have not been thrilled about some of the data you guys have used to train some of your models. Hollywood, publishers. As you work toward these next models, what are the conversations you’re having around the data?

ALTMAN: One, we obviously only want to use data that people are excited about us using.

But one of the challenges has been different kinds of data-owners have very different pictures. So we’re just experimenting with a lot of things. We’re doing partnerships of different shapes, and we think like with any new field, we’ll find something that sort of just becomes a new standard.

Also, as these models get smarter and more capable, we will need less training data. I think the conversation about data, and the shape of all of this, because of the technological progress we’re making, is about to shift.

WSJ: Well, publishers, like mine, they want money for that data. Is the future of this entire race about who can pay the most for the best data?

ALTMAN: No. That was sort of the point I was trying to make. The thing that people really like about a GPT model is not fundamentally that it knows particular knowledge. There are better ways to find that. It’s that it has this larval reasoning capacity, and that’s going to get better and better. That’s really what this is going to be about.

AI relations

WSJ: Is there a future where we have deep relationships with this type of bot?

MURATI: It’s going to be a significant relationship. Because we’re building the systems that are going to be everywhere: in your home, in your educational environment, in your work environment, and maybe when you’re having fun.

That’s why it’s actually so important to get it right. And we have to be so careful about how we design this interaction so that ultimately it’s elevating, fun, makes productivity better, and it enhances creativity.

We also want to make sure that, on the product side, we feel in control of these systems, in the sense that we can steer them to do the things that we want them to do and the output is reliable. As it has more information about your preferences, the things you like, the things you do, and as the capabilities of the models increase, it will become more personalized.

And it’s not just one system. You can have many such systems personalized for specific domains and tasks.

WSJ: That’s a big responsibility, though. And you will be sort of in control of people’s friends, maybe it gets to being people’s lovers. How do you think about that control?

ALTMAN: We’re not going to be the only player here. There’s going to be many people. So we get to put our “nudge” on the trajectory of this technological development. But: a) We really think that the decisions belong to humanity, society as a whole. And: b) We’ll be one of many actors building sophisticated systems here. There will be competing products that offer different things, there will be different kind of societal embraces and pushbacks, there will be regulatory stuff.

I personally have deep misgivings about this vision of the future where everyone is super close to AI friends, more so than human friends. I personally don’t want that. I accept that other people are going to want that, and some people are going to build that, and if that’s what the world wants, and what we decide makes sense, we’re going to get that.

The ‘fear’ issue

WSJ: We’ve got simple chatbots now. How do we go from that to this fear that is now pervading everywhere about AI?

ALTMAN: It doesn’t take much imagination to think about scenarios that deserve great caution. And, again, we all do this because we’re so excited about the tremendous upside and the incredibly positive impact. And it would be a moral failing not to pursue that for humanity. But we’ve got to address—and this happens with many other technologies—the downsides that come along with this.

It means that you are thoughtful about the risks. You try to measure what the capabilities are, and you try to build your own technology in a way that mitigates those risks. And then when you say like, “Hey, here’s a new safety technique,” you make that available to others.

WSJ: What are some of those?

MURATI: It’s not a single fix. You usually have to intervene everywhere, from the data to the model to the tools in the product, and, of course, policy. And then thinking about the entire regulatory and societal infrastructure that can keep up with these technologies that we’re building.

So, when you think about what are sort of the concrete safety measures along the way, No. 1 is actually rolling out the technology and slowly making contact with reality; understanding how it affects certain use-cases and industries; and actually dealing with the implications of that. Whether it’s regulatory, copyright, whatever the impact is, actually absorbing that, and dealing with that, and moving on to more and more capabilities. I don’t think that building the technology in a lab, in a vacuum—without contact with the real world and with the friction that you see with reality—is a good way to actually deploy it safely.

ALTMAN: The point Mira was making is really important. It’s very difficult to make a technology “safe” in the lab. Society uses things in different ways, and adapts in different ways. And I think the more we deploy AI, the more AI is used in the world, the safer AI gets, and the more we kind of collectively decide, “Hey, here’s a thing that is not an acceptable risk tolerance, and this other thing that people were worried about, that’s totally OK.”

Detection tools

WSJ: Right now there’s no tool, from OpenAI at least, that I can put in an image or some of the text, and ask, “Is this AI-generated?”, correct?

MURATI: For image, we have technology that’s almost 99% reliable. But we’re still testing it, it’s early.

WSJ: Is this something you plan to release?

MURATI: Yes.

WSJ: For both images and text?

MURATI: For text, we’re trying to figure out what makes sense. For images, it’s a bit more straightforward problem. Often we will experiment, we will put out something, we will get feedback. And sometimes we’ll take it back, make it better and roll it out again.

ALTMAN: This idea of “watermarking” content is not something everybody has the same opinion about what is good and what is bad. There are a lot of people who really don’t want their generated content watermarked. Also, it’s not going to be super-robust to everything. Like, maybe you could do it for images, and maybe for longer text, maybe not for short text.

This is why we want to engage in the conversation. We are willing to follow the collective wishes of society on this point. I don’t think it’s a black-and-white issue.

Work will change

AUDIENCE QUESTION: We’re about to change the nature of work. There’s a large portion of society that’s not even in this discussion. How do we come up with some of those frameworks and voluntarily bring things about that will actually result in a better world that doesn’t leave everybody else behind?

MURATI: I completely agree with you, that it’s the ultimate technology that could really increase inequality and make things so much worse for us as human beings and civilization. Or it could be amazing, and it could bring along a lot of creativity and productivity and enhance us. Maybe a lot of people don’t want to work eight hours, or 100 hours a week. Maybe they want to work four hours a day and do a bunch of other things.

It’s certainly going to lead to a lot of disruption in the workforce.

ALTMAN: It’s a super-important question. Every technological revolution affects the job market. I’m not afraid of that at all. In fact, I think that’s good, I think that’s the way of progress. And we’ll find new and better jobs.

The things that I think we do need to confront as a society is the speed at which this is going to happen. We are going to really have to do something about this transition. It is not enough to just give people a universal basic income. People need to have agency, the ability to influence this, we need to jointly be architects of the future.

As you said, not everybody’s in these discussions, but more are every year. And by putting this out in people’s hands, getting billions of people to use ChatGPT, not only do people have the opportunity to think about what’s coming and participate in that conversation, but people use the tool to push the future forward. And that’s really important to us.