Survival Strategies in the Era of AI Taught by Stanford

Li Jiang grew up inspired by Transformers to build robots, a passion that ultimately led him to direct the Stanford AIRE Program. Rather than fearing automation or banning tools like ChatGPT in classrooms, Jiang advocates for teaching AI thinking alongside Stanford's classic Design Thinking framework. His methodology trains students to delegate execution tasks to machine algorithms, allowing humans to focus on empathy, problem definition, and 0-to-1 innovation. In this interview, Jiang outlines how parents and educators can help students navigate artificial intelligence, re-architect learning assignments, and systematically build human creativity.

I'm Li Jiang, and I'm the director of the Stanford AIRE Program. The era of AI is in the near future, and I would say it's actually now. It has started already. I'm passionate about trying to figure out what the right way is of teaching the next generation in the era of AI and robotics.

A lot of people are afraid of AI, so they try to avoid it. They say they want to stay away from that. But the thing is that this is probably one of the most powerful tools we have ever invented. If you don't use these tools, then you cannot really compete with other people.

I think there are a few things we need to change. One is that we need to start to teach AI thinking as early as possible. AI thinking will allow them to know the difference between a human and an AI, and the human part is really the innovation part, from 0 to 1. That's where AI cannot do a good job. We need to focus on that.

There are 3 important things about AI thinking. The first thing is that you need to have a general understanding of how AI works. In the past, AI was mostly based on rules or specific algorithms that we write into the computer. The humans set the rules. Now we have developed different algorithms, like deep learning and reinforcement learning, based on data, because we have much more powerful computers that can handle way more data. They will try to find the most optimized solution based on the current data, and the more data you have, the better the solution you usually get.

If you do the first one, the second one is that you gain the ability to differentiate human ability from the machine's ability. I think probably most of the audience don't know there's a scientific research area called structural biology. Scientists in this field study the structure of a protein. Proteins are so important for all living creatures. The whole world has something like 100 to 200 million proteins.

In this research area, a lot of scientists focus on determining the structure of these proteins. We humans have found the structures of only a very small number of them, less than 1%. However, there is an AI called AlphaFold, from DeepMind. In the past two years, AlphaFold has predicted the structure of almost all the proteins in the world, and they put it in an open database online. So once the machines can do a job, let the machines do it, and then we focus on the human part.

After you get the first two, the third one is that you will have the ability to work with AI and use AI to help you accomplish other jobs. The critical part is going from 0 to 1, which basically means we invent something that didn't exist before, and we probably can do 0 to 1 better with AI providing a lot of information and things like that. So we should prepare our kids to learn how to innovate, to generate new ideas, and to use AI to help us with that kind of thing.

Creativity is something that is really hard to teach, and some people think creativity cannot be taught. How do you teach people creativity? This gets changed by Stanford design thinking. We can use this method to teach you how to systematically make innovations.

Design thinking is a methodology that we use here at Stanford to teach people how to make innovations. We do not make you as innovative or creative as Steve Jobs or Elon Musk. However, we can use this methodology to make you more creative and more innovative than yourself.

Design thinking has steps that you can follow. The first step is to empathize. First, you have to understand who you're inventing for. Are you inventing for your parents, or are you inventing for children who are studying in school? You have to understand who the users are, then you empathize with them, especially with their emotions.

Then you go to the second step, which is called define. A lot of people didn't have the right problem to start with. You have to define that problem clearly. Then you go to the third step, which is ideation, using brainstorming to generate a lot of ideas. You get feedback from the users and pick the good ones to prototype. After you do the prototype, which is the fourth step, you test with the user and get feedback. Maybe the feedback is good, maybe the feedback is bad, but you have to redesign or reinvent.

We have a phrase, all design is redesign. If you realize that you actually didn't define this problem right, then you go back to step 2. You have to go through these iterations. Eventually you will have some very nice inventions.

When you read books or listen to lectures, after you finish, you haven't understood it yet. You have to use it several times to really feel it. Let me give you an example. In one of the classes at Stanford, the project was about how to get more incubators for Nepal. At first, a lot of people were just trying to say, well, let's design it in California. But then we say, you have to gain empathy. You need to go to the real place. If you have never been to Nepal, how can you design for that?

They actually flew to Nepal and went to the mountains, and to their surprise, they saw a lot of these expensive incubators in the medical centers. They are not lacking those machines, but they don't know how to operate them. They had found the wrong problem. They had to redefine the problem, so they went back to these villages and talked with the farmers, who said, we do need those things, but not in the medical centers. We need them in our home, and we need the cheap ones.

So they designed a very cheap machine that functions as a baby incubator, and it had very good success. That's a good example. The problem was defined wrong at first, and then they went into this design thinking process, found a way to redefine the problem, and solved the right problem.

There are a few things that are most interesting about AI and robotics. I think the first one is ChatGPT, because it's probably a surprise for a lot of people that AI can talk with people in a very natural way. It cannot only chat with you, it can also write code for you. It seems like it knows a whole lot and is smarter than a lot of people, which surprised people. It's also going to change education for sure, because now kids can use ChatGPT to write essays and do homework. So teachers need to think about how to deal with that.

Actually, I just gave my students an assignment yesterday. I asked the students to write an essay. I said, we just talked about ChatGPT in the class, and now I'm going to give you an assignment to write an essay with the help of ChatGPT. I'm not asking you to not use it, I'm asking you to use it. And then after that class, I will have a short discussion with each of you about how you used it and how you feel about it.

I think that with technology advancement, you cannot try to ignore it or avoid it. It's happening, and it's just like water, you cannot block it. You have to go with it. For me, I want to understand how ChatGPT is going to impact the education system, and I want to work with our students to understand that better. That's why I gave this homework to my students.

I'm a big fan of Autobots and Transformers. I grew up with those cartoons and I'm a big fan. Even now, I'm still a big fan of Optimus Prime, and that really inspired me. I always wanted to make cool robots, from when I was a little kid, and I think most kids have a dream when they're young. It's easier to inspire a dream when someone is in the early ages. After people grow up, they kind of lose their interest. They lose their dream. They don't know what to do. It's actually critically important for these little minds to have dreams.

My ultimate goal is to find out what the ideal education system for the future is. The education part is really important. It has a profound impact on the next generation's future, and that is my ultimate goal.