We started OpenAI years ago because we felt like something really interesting was in AI and we wanted to help steer it in a positive direction. It’s honestly just amazing to see how far this whole field has come since then. it’s really gratifying to hear from people like Raymond are using the technology we are building, and others, for so many wonderful things. hear from people who are excited, we hear from people who concerned, we hear from people who feel both those emotions once. And honestly, that’s how we feel. Above all, feels like we’re entering an historic period right now where we as world are going to define a technology that will be so for our society going forward. And I believe that we manage this for good.
So today, I want to you the current state of that technology and some the underlying design principles that we hold dear.
So the thing I’m going to show you is what it’s like build a tool for an AI rather than building it for a human. So we have a DALL-E model, which generates images, and we are exposing it as an for ChatGPT to use on your behalf. And you can do like ask, you know, suggest a nice post-TED meal and a picture of it.
(Laughter)
Now you get all of the, sort of, ideation creative back-and-forth and taking care of the details for you you get out of ChatGPT. And here we go, it’s just the idea for the meal, but a very, detailed spread. So let’s see what we’re going to get. But ChatGPT doesn’t just generate images in this — sorry, it doesn’t generate text, it also generates an image. that is something that really expands the power of what it can on your behalf in terms of carrying out your intent. I’ll point out, this is all a live demo. This all generated by the AI as we speak. So actually don’t even know what we’re going to see. This looks wonderful.
(Applause)
I’m hungry just looking at it.
Now we’ve extended ChatGPT with other tools too, example, memory. You can say “save this for later.” the interesting thing about these tools is they’re very inspectable. So you get little pop up here that says “use the DALL-E app.” And by the way, this coming to you, all ChatGPT users, over upcoming months. And you look under the hood and see that what it did was write a prompt just like a human could. And so you sort of have this to inspect how the machine is using these tools, which allows us to provide to them.
Now it’s saved for later, and let me show you what it’s like use that information and to integrate with other applications too. You say, “Now make a shopping list for the tasty thing I suggesting earlier.” And make it a little tricky for the AI. “And tweet it for all the TED viewers out there.”
(Laughter)
So if you make this wonderful, wonderful meal, I definitely want to how it tastes.
But you can see that ChatGPT is selecting all these different tools me having to tell it explicitly which ones to use in any situation. And this, I think, a new way of thinking about the user interface. Like, are so used to thinking of, well, we have these apps, we click between them, copy/paste between them, and usually it’s a great experience an app as long as you kind of know the menus and know all the options. Yes, would like you to. Yes, please. Always good to be polite.
(Laughter)
And by this unified language interface on top of tools, the AI is able to sort take away all those details from you. So you don’t have to be the one who spells out single sort of little piece of what’s supposed to happen.
And I said, this is a live demo, so sometimes unexpected will happen to us. But let’s take a at the Instacart shopping list while we’re at it. And you can see we a list of ingredients to Instacart. Here’s everything you need. And the thing that’s really interesting is the traditional UI is still very valuable, right? If look at this, you still can click through it and sort of modify actual quantities. And that’s something that I think shows that they’re not away, traditional UIs. It’s just we have a new, augmented way to build them. And now have a tweet that’s been drafted for our review, which is a very important thing. We can click “run,” and there we are, we’re manager, we’re able to inspect, we’re able to change the of the AI if we want to. And so after this talk, will be able to access this yourself. And there we go. Cool. you, everyone.
(Applause)
So we’ll cut back to the slides. Now, the thing about how we build this, it’s not just about building tools. It’s about teaching the AI how to use them. Like, what we even want it to do when we ask these very high-level questions? to do this, we use an old idea. If you go back to Alan Turing’s 1950 on the Turing test, he says, you’ll never program answer to this. Instead, you can learn it. You could build a machine, a human child, and then teach it through feedback. a human teacher who provides rewards and punishments as it tries out and does things that are either good or bad.
And is exactly how we train ChatGPT. It’s a two-step process. First, we produce what Turing would have a child machine through an unsupervised learning process. We just show it the world, the whole internet and say, “Predict what comes next in text you’ve seen before.” And this process imbues it with all sorts of skills. For example, if you’re shown a math problem, only way to actually complete that math problem, to say what comes next, that green nine up there, to actually solve the math problem.
But we actually have to do second step, too, which is to teach the AI what to with those skills. And for this, we provide feedback. have the AI try out multiple things, give us multiple suggestions, then a human rates them, says “This one’s better than that one.” And this reinforces not the specific thing that the AI said, but very importantly, the whole that the AI used to produce that answer. And this allows it to generalize. It allows it teach, to sort of infer your intent and apply it in scenarios that it hasn’t before, that it hasn’t received feedback.
Now, sometimes the we have to teach the AI are not what you’d expect. For example, when we first GPT-4 to Khan Academy, they said, “Wow, this is great, We’re going to be able to teach students wonderful things. one problem, it doesn’t double-check students’ math. If there’s some bad in there, it will happily pretend that one plus one three and run with it.” So we had to collect feedback data. Sal Khan himself was very kind and 20 hours of his own time to provide feedback to machine alongside our team. And over the course of a couple of we were able to teach the AI that, “Hey, you really should push on humans in this specific kind of scenario.” And we’ve actually made lots and lots of to the models this way. And when you push that thumbs down in ChatGPT, that actually is of like sending up a bat signal to our team say, “Here’s an area of weakness where you should gather feedback.” so when you do that, that’s one way that really listen to our users and make sure we’re something that’s more useful for everyone.
Now, providing high-quality feedback is a hard thing. If you about asking a kid to clean their room, if all you’re doing is the floor, you don’t know if you’re just teaching them to stuff all the in the closet. This is a nice DALL-E-generated image, the way. And the same sort of reasoning applies AI. As we move to harder tasks, we will to scale our ability to provide high-quality feedback. But for this, the AI itself is happy help. It’s happy to help us provide even better feedback and to our ability to supervise the machine as time goes on. And me show you what I mean.
For example, you can ask GPT-4 a question like this, of much time passed between these two foundational blogs on unsupervised learning and learning from feedback. And the model says two months passed. But it true? Like, these models are not 100-percent reliable, although they’re getting every time we provide some feedback. But we can use the AI to fact-check. And it can actually check its own work. You can say, fact-check this me.
Now, in this case, I’ve actually given the AI a new tool. This one is a tool where the model can issue search queries and click into web pages. And it actually writes its whole chain of thought as it does it. It says, I’m just going to search for this it actually does the search. It then it finds the publication date the search results. It then is issuing another search query. It’s going to click into the blog post. And all this you could do, but it’s a very tedious task. It’s a thing that humans really want to do. It’s much fun to be in the driver’s seat, to be this manager’s position where you can, if you want, triple-check work. And out come citations so you can actually go and easily verify any piece of this whole chain of reasoning. it actually turns out two months was wrong. Two and one week, that was correct.
(Applause)
And we’ll back to the side. And so thing that’s so interesting to me this whole process is that it’s this many-step collaboration between human and an AI. Because a human, using this fact-checking is doing it in order to produce data for another AI to become more useful to a human. I think this really shows the shape of something that should expect to be much more common in the future, where we have humans and machines kind of very and delicately designed in how they fit into a problem and how we to solve that problem. We make sure that the humans providing the management, the oversight, the feedback, and the are operating in a way that’s inspectable and trustworthy. together we’re able to actually create even more trustworthy machines. And I think that over time, if we get process right, we will be able to solve impossible problems.
And to you a sense of just how impossible I’m talking, I think we’re going be able to rethink almost every aspect of how interact with computers. For example, think about spreadsheets. They’ve around in some form since, we’ll say, 40 years ago VisiCalc. I don’t think they’ve really changed that much in time. And here is a specific spreadsheet of all the papers on the arXiv for the past 30 years. There’s 167,000 of them. And you can see there the data here. But let me show you the ChatGPT take on how analyze a data set like this.
So we can ChatGPT access to yet another tool, this one a Python interpreter, it’s able to run code, just like a data scientist would. And you can just literally upload a file and ask about it. And very helpfully, you know, it knows the name of the file it’s like, “Oh, this is CSV,” comma-separated value file, “I’ll it for you.” The only information here is the name the file, the column names like you saw and the actual data. And from that it’s able to infer what columns actually mean. Like, that semantic information wasn’t in there. has to sort of, put together its world knowledge knowing that, “Oh yeah, arXiv is a site that people papers and therefore that’s what these things are and that are integer values and so therefore it’s a number of authors in the paper,” like of that, that’s work for a human to do, and the AI is happy help with it.
Now I don’t even know what I to ask. So fortunately, you can ask the machine, “Can make some exploratory graphs?” And once again, this is a high-level instruction with lots of intent behind it. But I don’t even know what want. And the AI kind of has to infer I might be interested in. And so it comes with some good ideas, I think. So a histogram the number of authors per paper, time series of per year, word cloud of the paper titles. All that, I think, will be pretty interesting to see. And the great thing is, it can actually it. Here we go, a nice bell curve. You see that three is kind the most common. It’s going to then make this nice plot of papers per year. Something crazy is happening in 2023, though. Looks like we were on an exponential and it off the cliff. What could be going on there? By the way, all this is code, you can inspect. And then we’ll see word cloud. So you can see all these wonderful that appear in these titles.
But I’m pretty unhappy about this 2023 thing. makes this year look really bad. Of course, the problem that the year is not over. So I’m going to push back on machine. [Waitttt that’s not fair!!! 2023 isn’t over. What percentage of papers in 2022 were even by April 13?] So April 13 was the cut-off date believe. Can you use that to make a fair projection? we’ll see, this is the kind of ambitious one.
(Laughter)
So know, again, I feel like there was more I wanted out of the machine here. really wanted it to notice this thing, maybe it’s little bit of an overreach for it to have of, inferred magically that this is what I wanted. But inject my intent, I provide this additional piece of, you know, guidance. under the hood, the AI is just writing code again, if you want to inspect what it’s doing, it’s very possible. And now, it the correct projection.
(Applause)
If you noticed, it even updates the title. I didn’t ask that, but it know what I want.
Now we’ll cut back to slide again. This slide shows a parable of how I think we … vision of how we may end up using this technology in the future. person brought his very sick dog to the vet, and the veterinarian a bad call to say, “Let’s just wait and see.” And the dog would be here today had he listened. In the meanwhile, he provided blood test, like, the full medical records, to GPT-4, said, “I am not a vet, you need to talk a professional, here are some hypotheses.” He brought that information to a second vet who it to save the dog’s life. Now, these systems, they’re perfect. You cannot overly rely on them. But this story, I think, shows a human with a medical professional and with ChatGPT as a brainstorming was able to achieve an outcome that would not have otherwise. I think this is something we should all on, think about as we consider how to integrate these systems our world.
And one thing I believe really deeply, is that getting AI right going to require participation from everyone. And that’s for deciding we want it to slot in, that’s for setting the rules of the road, what an AI will and won’t do. And if there’s one thing take away from this talk, it’s that this technology just different. Just different from anything people had anticipated. And we all have to become literate. And that’s, honestly, one the reasons we released ChatGPT.
Together, I believe that we achieve the OpenAI mission of ensuring that artificial general intelligence all of humanity.
Thank you.
(Applause)
(Applause ends)
Chris Anderson: Greg. Wow. I mean … I suspect that within every out here there’s a feeling of reeling. Like, I that a very large number of people viewing this, you look at that and think, “Oh my goodness, pretty much every single thing about way I work, I need to rethink.” Like, there’s new possibilities there. Am I right? Who thinks that they’re having to rethink the way we do things? Yeah, I mean, it’s amazing, but it’s really scary. So let’s talk, Greg, let’s talk.
I mean, I guess first question actually is just how the hell have done this?
(Laughter)
OpenAI has a few hundred employees. Google has thousands of employees working on intelligence. Why is it you who’s come up with technology that shocked the world?
Greg Brockman: I mean, the is, we’re all building on shoulders of giants, right, there’s no question. If you look at compute progress, the algorithmic progress, the data progress, all those are really industry-wide. But I think within OpenAI, we made lot of very deliberate choices from the early days. And first one was just to confront reality as it lays. And that we just thought really hard about like: is it going to take to make progress here? We tried a lot things that didn’t work, so you only see the things that did. And I that the most important thing has been to get teams people who are very different from each other to together harmoniously.
CA: Can we have the water, by way, just brought here? I think we’re going to it, it’s a dry-mouth topic. But isn’t there something also about the fact that you saw something in these language models that meant that you continue to invest in them and grow them, that something at point might emerge?
GB: Yes. And I think that, I mean, honestly, I the story there is pretty illustrative, right? I think that high level, deep learning, like we always that was what we wanted to be, was a deep learning lab, and how to do it? I think that in the early days, didn’t know. We tried a lot of things, and one person was working on a model to predict the next character in Amazon reviews, and got a result where — this is a syntactic process, expect, you know, the model will predict where the go, where the nouns and verbs are. But he got a state-of-the-art sentiment analysis classifier out of it. This model tell you if a review was positive or negative. I mean, today we are just like, on, anyone can do that. But this was the first time that you saw this emergence, sort of semantics that emerged from this underlying syntactic process. there we knew, you’ve got to scale this thing, you’ve got to see where goes.
CA: So I think this helps explain the riddle that baffles everyone looking at this, these things are described as prediction machines. And yet, what we’re seeing out of them feels … just feels impossible that that could come from a machine. Just the stuff you showed us just now. And key idea of emergence is that when you get of a thing, suddenly different things emerge. It happens the time, ant colonies, single ants run around, when you bring enough of them together, you get these colonies that show completely emergent, different behavior. Or a where a few houses together, it’s just houses together. as you grow the number of houses, things emerge, suburbs and cultural centers and traffic jams. Give me one moment for you when saw just something pop that just blew your mind that just did not see coming.
GB: Yeah, well, so can try this in ChatGPT, if you add 40-digit —
CA: 40-digit?
GB: 40-digit numbers, the model will do it, which it’s really learned an internal circuit for how to do it. And the interesting thing is actually, if you have it add a 40-digit number plus a 35-digit number, it’ll often get it wrong. And so you can see it’s really learning the process, but it hasn’t fully generalized, right? It’s you can’t memorize the 40-digit addition table, that’s more atoms there are in the universe. So it had to have learned general, but that it hasn’t really fully yet learned that, Oh, I can sort of generalize this to adding arbitrary of arbitrary lengths.
CA: So what’s happened here is that you’ve allowed it to up and look at an incredible number of pieces text. And it is learning things that you didn’t that it was going to be capable of learning.
GB Well, yeah, it’s more nuanced, too. So one science that we’re starting to really good at is predicting some of these emergent capabilities. And to do that actually, one of the I think is very undersung in this field is sort engineering quality. Like, we had to rebuild our entire stack. When think about building a rocket, every tolerance has to be incredibly tiny. Same is true machine learning. You have to get every single piece of stack engineered properly, and then you can start doing predictions. There are all these incredibly smooth scaling curves. tell you something deeply fundamental about intelligence. If you look our GPT-4 blog post, you can see all of these curves in there. now we’re starting to be able to predict. So we were able to predict, example, the performance on coding problems. We basically look at models that are 10,000 times or 1,000 times smaller. And there’s something about this that is actually smooth scaling, even it’s still early days.
CA: So here is, one of the big then, that arises from this. If it’s fundamental to what’s happening here, that as you scale up, things emerge you can maybe predict in some level of confidence, it’s capable of surprising you. Why isn’t there just huge risk of something truly terrible emerging?
GB: Well, I think all these are questions of degree and scale and timing. And I think one thing people miss, too, sort of the integration with the world is also incredibly emergent, sort of, very powerful thing too. And that’s one of the reasons that we think it’s so important to deploy incrementally. so I think that what we kind of see now, if you look at this talk, a lot of what I on is providing really high-quality feedback. Today, the tasks we do, you can inspect them, right? It’s very easy to at that math problem and be like, no, no, no, machine, seven the correct answer. But even summarizing a book, like, that’s a hard to supervise. Like, how do you know if this book summary is any good? You have to read whole book. No one wants to do that.
(Laughter) And so I think that the important thing will that we take this step by step. And that say, OK, as we move on to book summaries, we have to this task properly. We have to build up a record with these machines that they’re able to actually carry out our intent. And think we’re going to have to produce even better, efficient, more reliable ways of scaling this, sort of like making the be aligned with you.
CA: So we’re going to hear in this session, there are critics who say that, know, there’s no real understanding inside, the system is to always — we’re never going to know that it’s generating errors, that it doesn’t have common sense and so forth. it your belief, Greg, that it is true at any one moment, that the expansion of the scale and the human feedback that you talked is basically going to take it on that journey of actually getting to things like and wisdom and so forth, with a high degree confidence. Can you be sure of that?
GB: Yeah, well, I that the OpenAI, I mean, the short answer is yes, I that is where we’re headed. And I think that the approach here has always been just like, let reality you in the face, right? It’s like this field is field of broken promises, of all these experts saying is going to happen, Y is how it works. People have been saying neural aren’t going to work for 70 years. They haven’t been right yet. They might right maybe 70 years plus one or something like is what you need. But I think that our has always been, you’ve got to push to the of this technology to really see it in action, because tells you then, oh, here’s how we can move on to a new paradigm. we just haven’t exhausted the fruit here.
CA: I mean, it’s a controversial stance you’ve taken, that the right way to do this is to it out there in public and then harness all this, you know, of just your team giving feedback, the world is now giving feedback. But … If, know, bad things are going to emerge, it is there. So, you know, the original story that I heard on OpenAI when you were founded as nonprofit, well you were there as the great sort check on the big companies doing their unknown, possibly evil thing with AI. And you were going build models that sort of, you know, somehow held them and was capable of slowing the field down, if need be. Or at that’s kind of what I heard. And yet, what’s happened, arguably, is the opposite. That your of GPT, especially ChatGPT, sent such shockwaves through the tech world that now Google and Meta so forth are all scrambling to catch up. And of their criticisms have been, you are forcing us put this out here without proper guardrails or we die. You know, do you, like, make the case that what you have done is responsible and not reckless.
GB: Yeah, we think about these questions the time. Like, seriously all the time. And I don’t we’re always going to get it right. But one thing I has been incredibly important, from the very beginning, when we were about how to build artificial general intelligence, actually have it all of humanity, like, how are you supposed to do that, right? And that plan of being, well, you build in secret, you get this powerful thing, and then you figure out the safety it and then you push “go,” and you hope you it right. I don’t know how to execute that plan. Maybe else does. But for me, that was always terrifying, didn’t feel right. And so I think that this alternative approach is the only other path I see, which is that you do let reality hit in the face. And I think you do give people time to give input. You do have, before machines are perfect, before they are super powerful, that you actually have ability to see them in action. And we’ve seen from GPT-3, right? GPT-3, we really were afraid that the number one thing people were going to do it was generate misinformation, try to tip elections. Instead, the number one thing was generating spam.
(Laughter)
CA: So Viagra spam is bad, but are things that are much worse. Here’s a thought for you. Suppose you’re sitting in a room, there’s a box on the table. believe that in that box is something that, there’s a very strong chance it’s something absolutely glorious that’s to give beautiful gifts to your family and to everyone. But there’s actually a one percent thing in the small print there that says: “Pandora.” there’s a chance that this actually could unleash unimaginable evils on the world. Do you that box?
GB: Well, so, absolutely not. I think you don’t do it that way. honestly, like, I’ll tell you a story that I haven’t told before, which is that shortly after we started OpenAI, I remember I was in Puerto Rico an AI conference. I’m sitting in the hotel room looking out over this wonderful water, all these people having a good time. And you about it for a moment, if you could choose for basically that Pandora’s to be five years away or 500 years away, which would pick, right? On the one hand you’re like, well, for you personally, it’s better to have it be five away. But if it gets to be 500 years and people get more time to get it right, which do you pick? And you know, I really felt it in the moment. I was like, of course you the 500 years. My brother was in the military at the time like, he puts his life on the line in a much more real way any of us typing things in computers and developing technology at the time. And so, yeah, I’m really on the you’ve got to approach this right. But I don’t that’s quite playing the field as it truly lies. Like, if you look at the whole of computing, I really mean it when I say that this is an industry-wide even just almost like a human-development- of-technology-wide shift. And the that you sort of, don’t put together the pieces that there, right, we’re still making faster computers, we’re still the algorithms, all of these things, they are happening. And you don’t put them together, you get an overhang, which means if someone does, or the moment that someone does manage connect to the circuit, then you suddenly have this very powerful thing, no one’s had any time adjust, who knows what kind of safety precautions you get. so I think that one thing I take away is like, even think about development of other sort of technologies, think about nuclear weapons, people talk about like a zero to one, sort of, change in what humans could do. But I actually think if you look at capability, it’s been quite smooth over time. And so the history, I think, of technology we’ve developed has been, you’ve got to do it incrementally and you’ve to figure out how to manage it for each moment you’re increasing it.
CA: So what I’m hearing is that you … the model you want us have is that we have birthed this extraordinary child that may have superpowers that take humanity to a new place. It is our collective responsibility to provide the guardrails this child to collectively teach it to be wise and not to us all down. Is that basically the model?
GB: think it’s true. And I think it’s also important say this may shift, right? We’ve got to take step as we encounter it. And I think it’s incredibly important today that we all do literate in this technology, figure out how to provide feedback, decide what we want from it. And my is that that will continue to be the best path, it’s so good we’re honestly having this debate because we wouldn’t otherwise if weren’t out there.
CA: Greg Brockman, thank you so much for coming to TED blowing our minds.
(Applause)