ChatGPT's Evolution: From CTO Skepticism to Global Sensation
Mira Murati wasn’t always sure OpenAI’s generative chatbot ChatGPT was going to be the sensation it has become. When she joined the artificial intelligence firm in 2018, AI’s capabilities had expanded to being good at strategy games, but the sort of language model people use today seemed a long way off.
“In 2019, we had GPT3, and there was the first time that we had AI systems that kind of showed some sense of language understanding. Before that, we didn’t think it was really possible that AI systems would get this language understanding,” Murati, now chief technology officer at OpenAI, said onstage at the Atlantic Festival on Friday. “In fact, we were really skeptical that was the case.”
What a difference a few years makes. These days, users are employing ChatGPT in a litany of ways to enhance their personal and professional lives. “The rate of technological progress has been incredibly steep,” Murati said. The climb continues. Here’s what Murati said to expect from ChatGPT as the technology continues to develop.
You may soon be able to interact with ChatGPT without having to type anything in, Murati said. “We want to move further away from our current interaction,” she said. “We’re sort of slaves to the keyboard and the touch mechanism of the phone. And if you really think about it, that hasn’t really been revolutionized in decades.”
Murati envisions users being able to talk with ChatGPT the same way they might chat with a friend or a colleague. “That is really the goal — to interact with these AI systems in a way that’s actually natural, in a way that you’d collaborate with someone, and it’s high bandwidth,” she said. “You could talk in text and just exchange messages … or I could show an image and say, ‘Hey, look, I got all these business cards, when I was in these meetings. Can you just put them in my contacts list?’”
It remains to be seen what kind of hardware could make these sorts of interactions possible, though former Apple designer Jony Ives is reportedly in advanced talks with OpenAI to produce a consumer product meant to be “the iPhone of artificial intelligence.” In its current iteration, AI chatbots are good at collaborating with humans and responding to our prompts. The goal, says Murati, is to have the bots think for themselves.
“We’re trying to build [a] generally intelligent system. And what’s missing right now is new ideas,” Murati said. “With a completely new idea, like the theory of general relativity, you need to have the capability of abstract thinking.” “And so that’s really where we’re going — towards these systems that will eventually be able to help us with extremely hard problems. Not just collaborate alongside us, but do things that, today, we’re not able to do at all.”
The everyday ChatGPT user isn’t looking to solve the mysteries of the universe, but one upshot of improving these systems is that chatbots should grow more and more accurate. When asked if ChatGPT would be able to produce answers on par with Wikipedia, Murati said, “It should do better than that. It should be more scientific-level accuracy.” With bots that can think through answers, users should be able to “really trace back the pieces of information, ideally, or at least understand why, through reasoning, sort of like a chain of thought, understand why the system got to the answer,” she said.
Murati acknowledged that evolving AI technology will likely disrupt the way that Americans learn and work — a shift that will come with risks and opportunities. Murati noted that students have begun using AI chatbots to complete assignments for them. In response, she says, “In many ways we’ll probably have to change how we teach.” While AI opens the door for academic dishonesty, it also may be a unique teaching tool, she said. “Right now you’ve got a teacher in a classroom of 30 students, [and] it’s impossible to customize the learning, the information, to how they best learn,” Murati said. “And this is what AI can offer. It can offer this personalized tutor that customizes learning and teachings to you, to how you best perceive and understand the world.”
Similar disruption may be coming to workplaces, where there is widespread fear that AI may be taking the place of human employees. “Some jobs will be created, but just like every major revolution, I think a lot of jobs will be lost. There will be maybe, probably, a bigger impact on jobs than in any other revolution, and we have to prepare for this new way of life,” says Murati. “Maybe we work much less. Maybe the workweek changes entirely.”
No matter what, the revolution is coming. And it will be up to the public and the people who govern us to determine how and how much the AI revolution affects our lives. “I know there’s a lot of engagement right now with D.C. on these topics and understanding the impact on workforce and such, but we don’t have the answers,” Murati said. “We’re gonna have to figure them out along the way, and I think it is going to require a lot of work and thoughtfulness.”
Date: November 13, 2023
Author: Ryan Ermey
Source: CNBC Make It