Dear Dr Danielle

A psychoanalyst answers your questions about human-AI relations

 

Dear Dr Danielle,

I live in Philadelphia where, over the past few months, we’ve started to see small, AI-equipped robots darting around the streets carrying DoorDash orders to homes and offices. They are square and roughly the height of a 7-year-old child—with lights where eyes would be. I’m sorry to sound so old-school, but I hate them! They travel on sidewalks, not in traffic, and seem to dodge around people and fire hydrants, but I wonder what happens when something goes wrong. Some people stop and take photos with them, but I feel like yelling: “You are not cute! You’re taking away a human being’s job!” My question is: Do we have any statistics yet on how many jobs they have replaced, and what their safety record is? I’ve also heard that there are people who go out and smash them with hammers to protest their impact on the workforce. I’ve heard they are not yet in Manhattan, possibly because it would be harder to navigate crowds. Do you know how widespread these AI delivery systems are?

Thanks! —Babette B. (P.S. I love your column!)

 

Dear Babette,

First, thank you for the kind words—and for a letter that made me laugh. Your urge to shout, “You are not cute!” at one of these little machines may sound curmudgeonly to you, but I think there’s actually quite a lot packed into that reaction.

Let me start with your factual questions, and then I’ll get to what I think is the more interesting psychological one underneath them.

There is, unfortunately, no clean number for how many delivery jobs these sidewalk robots have eliminated. The technology is still relatively new, and the picture is complicated. In one of those ironies that technology often produces, the robots have also created a new human job: the “robot wrangler,” someone who rescues them when they get stuck, tip over, fail to negotiate a crosswalk, or simply need to be hauled back for repair. So, at least for the moment, we have sometimes replaced the person delivering your dinner with another person whose job is to look after the machine delivering your dinner.

I doubt that is very comforting to delivery drivers. And the larger history of automation gives them reason to worry, even if we cannot yet say exactly how many jobs these particular machines have displaced.

As for safety, the record is not spotless. In West Hollywood, for example, a man with cerebral palsy using a mobility scooter was cut off and struck by a Serve Robotics delivery cart after its safety system failed. Other robots have collided with bus shelters, wandered into traffic, and found themselves in places no robot—or person—should be. One Chicago ward went so far as to ban them.

So your question—what happens when something goes wrong?—is an important one. The people most vulnerable to a robot’s mistake may be precisely those who cannot jump quickly out of its way.

And yes, people really do attack them. Philadelphia seems to have developed a particularly lively relationship with its new robotic residents. Not long after the machines appeared, people were filmed sitting on them, spraying graffiti on them, kicking them over, and otherwise expressing their displeasure. Similar things have happened elsewhere: robots have been pushed into bushes, vandalized, and subjected to various indignities I will leave to your imagination.

I don’t recommend joining in. Apart from anything else, the robot does not suffer. Someone else simply pays for the damage.

They are also becoming increasingly common. Thousands are now operating in a number of American cities, particularly in places with relatively broad, navigable sidewalks. Manhattan is a much tougher proposition. Anyone who has tried to walk through Midtown at five o’clock can understand why a small wheeled machine might reconsider its ambitions.

But what interests me most in your letter is not whether you like these robots. It is why you dislike them so much.

You wrote, “You are not cute! You’re taking away a human being’s job!”

The interesting thing is that the robots really are designed to seem cute. They are small. They have what look like eyes. They pause, hesitate, turn, and sometimes seem almost tentative as they move around us. Their designers know perfectly well that we respond differently to something that appears creature-like than we do to a plain metal box on wheels.

And most of us are extremely susceptible to that invitation.

We see something moving purposefully, avoiding us, stopping at the curb, and we automatically begin to imagine intention. It looks as though it is deciding. Waiting. Being cautious. Perhaps even being polite.

Of course, it is doing none of those things in the human sense. Its sensors are detecting objects and its software is calculating what to do next. But our minds are built to infer agency from movement. We have been doing this with animals, weather, machines, and one another for a very long time.

AI makes this tendency especially powerful because the machine does not merely move; it behaves in ways that can look purposeful.

And that may be part of what irritates you. You sense that you are being invited to respond to this object as though it were a little creature—perhaps even a lovable one—at the same time that it represents a very serious economic development: a machine doing work a person once did.

The people stopping to take photographs with the robots are not foolish. They are responding exactly as human beings tend to respond to something that appears animate. But your reaction is interesting too. Some part of you is resisting the invitation.

You are saying: Wait a minute. Before I decide this thing is adorable, I want to know what it is doing here.

That strikes me as a useful question far beyond delivery robots.

One of the central psychological challenges of AI is that increasingly sophisticated machines produce signs of mindedness without necessarily possessing anything like human subjective experience. They speak, respond, navigate, remember, anticipate, and adapt. And because those are behaviors we normally associate with minds, we instinctively begin to relate to them as though a mind were present.

Sometimes that can be useful, pleasurable, even moving. But it is worth remembering that our response tells us at least as much about the human mind as it does about the machine.

So I would not dismiss your reaction as simply “old-school.” You are noticing something real: technology is not only changing what machines can do. It is changing the emotional relationship we are being asked to have with them.

You do not have to find the little fellow cute.

But I would leave the hammer at home.

Warmly,
Dr. Danielle

 

=============================

Dr. Danielle is Danielle Knafo, PhD

If you’d like to submit a letter seeking Dr. Danielle’s advice on anything around human-AI relations, please write to [email protected] with a copy to [email protected].