Skip to Content
Smart cities

A day in the life of a Chinese robotaxi driver

We spoke to Liu Yang, who has one of the strangest jobs around: to sit in the passenger seat and monitor how self-driving cars cope with Beijing’s streets.

July 27, 2022
baidu worker (left) and autonomous vehicle driving on highway (right)
Courtesy of Baidu

When Liu Yang started his current job, he found it hard to go back to driving his own car. “I instinctively went for the passenger seat. Or when I was driving, I would expect the car to brake by itself,” says the 33-year-old Beijing native, who joined the Chinese tech giant Baidu in January 2021 as a robotaxi driver.

If you live in a place like San Francisco, Phoenix, or Shenzhen, you are probably no stranger to the self-driving taxis that roam the city. But the robotaxis are often still “drivered.” For example, the Chinese government requires a person to be present in the car, even if not controlling how the car moves. Liu is one of the hundreds of “drivers”—or safety operators, as they are called in the industry—that Baidu employs.

Robotaxi safety operator is an occupation that exists only in our time, the result of an evolving technology that’s advanced enough to get rid of a driver—most of the time, and in controlled environments—but not good enough to convince authorities that they can do away with human intervention altogether. Today, self-driving companies from the US, Europe, and China are racing to bring the technology to commercial application. Most of them, including Apollo, the self-driving arm of Baidu, have started on-demand robotaxi trials on public roads but still need to operate with various constraints.

With an associate degree in human resources, Liu has no academic training related to this job, But he has always loved driving, and he acted as the driver for his boss in a previous role. When he heard about the self-driving technologies, his curiosity pushed him to look up related jobs online and apply. Today Liu, who has a buzz cut, a warm smile, and a distinctive Beijing accent, “drives” a robotaxi five days a week in Shougang Park, a 3.3-square-mile former power plant in Beijing that has been redeveloped into a tourist attraction after serving as a sports venue for the 2022 Winter Olympics. His car can’t leave the park, which was designated as a trial area for robotaxis, so his passengers are usually employees who work there or tourists visiting on weekends.

But Liu also needs to think about his next steps, as his job will likely be eliminated within a few years. He has been through several robotaxi models and policy changes in his 19-month career as a safety operator. In April 2021, Baidu acquired the license to put the safety operator in the front passenger seat instead of the driver’s seat (only within Shougang Park), and Liu subsequently switched his position and said goodbye to the steering wheel. On July 21 this year, Baidu revealed its new robotaxi model whose steering wheel can be removed. It is expected to be in operation in 2023.

MIT Technology Review talked to Liu Yang in June. We asked him about how he came to get this job, what his daily life is like, and what the future holds for an occupation designed to disappear soon.

The interview has been translated from Chinese and edited for clarity.

How did you decide to become a safety operator for a self-driving taxi?

It was quite a coincidence. Back when I was parking the car for an old boss, I didn’t know what self-driving was and saw that his car had a self-parking function. I was super, super curious. It was really interesting when we ordinary people try it for the first time. After that, I wanted to know more.

How long have you been driving?

Twelve, 13 years.

Did you remember what the interview process was like?

I was extremely nervous when I went for the interview. We had two rounds, a face-to-face interview and a road test. I think the road test for self-driving [operators] is more difficult than the road test for getting a driver’s license. When you are learning how to drive, you need to look at your left, your right, and the rearview mirror; but when we are taking the test [for Baidu], you need to pay attention to all the directions, as well as what every car in front of or behind you is doing. Maybe it will suddenly change lanes and impact you.

What is your daily schedule like as a safety operator?

Work starts at 10 a.m.and lasts till 6 p.m. I spend on average seven hours in the car every day. Since we are commercial vehicles, we need to make sure there are always cars available when the passengers order them. So we [drivers] take turns to have lunch, usually between 11 a.m. and 1 p.m. If there have been a lot of rides that day and the battery needs to be recharged, we would recharge the car when we eat lunch.

What are you responsible for when you are in a car, since you don’t actually drive it?

When I’m in the driver’s seat, I pay attention to the road and then steer the car if there’s any problem. When I’m in the passenger seat, I pay more attention to the surroundings of the car, like through the rearview mirror. Because I don’t have access to the steering wheel, I may talk to the cloud-based operators [if a human needs to intervene].

(Editor’s note: Baidu Apollo has cloud-based remote human operators ready to assume control over a robotaxi in case of emergency.)

What did it feel like when you first started the job?

Because I’d never been in a self-driving car, I would get afraid when I first started this job because I was not the one controlling it. Like when it was making a turn, maybe I [as a driver] would turn this wide, but it would turn that wide. The car will follow the legally required paths, but when we are driving ourselves, we may occasionally cut corners or cross the solid line. So the paths we would take are different. I did have some fear, but it was actually safe. It drives itself according to the traffic laws.

Now that you’ve been working as a safety operator for a year and a half, does it feel weird going back to drive your own car?

I couldn’t get used to it in the beginning. After a day’s work was over and when I got back in my own car, I would treat it as if it were a self-driving car. Sometimes, when I got to my car I instinctively went for the passenger seat. Or when I was driving, I would expect the car to brake by itself. And then I realized that’s not right. It’s me [driving this car.] But I got used to it after a few times, and then it’s back to normal.

Ever since you started to have passengers in your car, has your work experience changed?

In the past it was just me. I drove the car, examined the car all by myself for the whole day. But now since we’ve started the commercial operation, I would be meeting different people every day, and the variety of passengers makes the work less dry than before. Because everyone is curious about self-driving, they want to chat with you and ask you questions.

How many passengers do you have every day?

On average 10 to 15 rides. And then it’s usually one or two passengers each ride. The rides last about four or five minutes.

Has anyone ever doubted the safety of the car and wanted you to drive it manually?

No, but there are many who are interested in the car and want to drive it themselves. They would ask if they could sit in the driver’s seat, and I’d say no.

Are there a lot of people asking to be in the driver’s seat?

Thirty percent.

Since you are confined to the Shougang Park every day, have you learned the area very well, or are you getting bored with it?

I wouldn’t say I’m bored, but I know the streets in Shougang very well. I know at which crossroads there’s likely to be a blind spot for normal cars to interfere with our work, or where pedestrians are more likely to jaywalk. I probably know more about those than people visiting Shougang.

Like there’s one crossroad in Shougang where there are some plants dividing the roads. That area is a major blind spot. When cars come in from outside, the view will be blocked [by the vegetation], so sometimes you will have trouble with oncoming traffic. But for me, after we’ve been working here for a while, we know well when we approach there and will observe the situation in advance to prevent accidents from happening.

How do you feel about your job? Do you feel like you are having an impact on the progression of self-driving technology?

Since we are in the car every day, we can identify more problems and report them afterwards. We help the technology departments keep up with what’s happening to the cars on a daily basis and what problems need to be solved.

The goal with self-driving is to eliminate any driver or safety operator. Let‘s say in one year or five years, no safety operator is needed anymore. What’s your career plan?

I want to stay in this industry. I have high hopes for it, and I like it a lot. I entered this industry because I was curious, and now I still believe it’s advanced and powerful. I think it’s very futuristic.

If I can have more time to improve myself, I want to learn more on the technical side, like data technologies. The other thought I had was, with the knowledge and the experience I have from doing this job for so long, I can train new people, people in other sectors, to learn about self-driving. I still want to stay in self-driving.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.