Updated Jan. 26, 2018
A professor at Stanford University and author of The Man Who Lied to His Laptop, Clifford Nass is an expert on how people respond to technology. His experiments with humans and computers show why traditional workplace strategies for managing employees don’t work.
Q: How did you come to the idea that people can’t emotionally distinguish between computers and fellow humans?
A: I had noticed how people’s responses to the Microsoft animated paperclip Clippy were so strong and vehement, and how German male drivers wouldn’t “take directions from a woman” even though it was obviously only a woman’s voice in the car’s computer. I was really stuck trying to figure out these responses to machines, and then I realized the computers were unconsciously viewed as being alive. I began experiments just trying to understand all the ways that people do treat computers like people, and it turned out that it’s in everything from being polite to accepting flattery.
Q: Did you start with the politeness test?
A: Yes, that was the first, most surprising and most powerful result. If you ask me how much I like your writing I’m obviously going to say nicer things to you than I would to some third party, no matter how much in fact I do like your writing. And the reason is we don’t want to hurt people’s feelings, so the question was, even if we know that a computer does not have feelings, nonetheless would we be polite to it? So we asked people to evaluate a computer’s performance, either on that particular computer or on another one across the room. And to our shock, people said nicer things to the computer being evaluated than they did to the computer across the room. These were not novices who might not understand that computers don’t have feelings, these were computer science grad students at Stanford. And in fact they insisted they paid no attention to anything so crazy. But they did and that was pretty compelling.
Q: That we would feel reciprocity toward a computer that helped us is fascinating.
A: Yeah, that one is pretty startling, that if a computer is helpful to a user and then asks the user for help, it gets help far more often than a computer that wasn’t helpful itself. It’s actually a great example of these principles. Humans are built to be relatively simple, even though they seem so complicated, and reciprocity is one of those principles we seem designed for. Once something cues that stimulus, it makes us feel, “Wow, I owe a favour,” and then the entire cognitive apparatus leaps in to deliver. We tend not to notice because we’re living our lives, and we don’t get to act and observe.
Q: Does that mean that acting badly requires more cognitive power, actual thought, while returning a favour needs no thought at all?
A: Negative things take up more cognitive energy: we think harder about them, we feel them more strongly. It’s why criticism is so much more powerful than praise. It’s true that it takes much more attention to do something negative like criticize as opposed to praise. And once you get on a roll in criticizing it comes pouring out, whereas with praise we don’t see that. We don’t hear about bursts of kindness, we have bursts of anger.
Q: The phrases are more limited, too.
A: That’s right: 50 per cent of all words are negative, and only 30 per cent are positive, in both English and Spanish. So we tend to have a much more colourful, rich, negative vocabulary, and it’s all because our brains are built to be particularly excited by negative things.
Q: So a computer, by seeming slightly human—responding to queries, say—triggers automatic human-to-human responses in us?
A: I used to be in the magician’s union, which included ventriloquists, and I was talking to one who said, “You know, as a magician you get all the credit, right? But as a ventriloquist the star of the show is the dummy.” So I said, “Why don’t you make it clear that you’re the star by going out there with just an artichoke on your lap? Do the show, and then people will realize that you are the incredibly talented one.” So I convinced this guy to do it, and the audience hated it: booing, screaming. I asked what was the problem, and he says, “Artichokes can’t talk.” But the audience would have been willing to accept that a hunk of wood could talk. Because it had a face!
Q: Okay, I get the wood in human form and the sensitive computer, but the TV experiment must have surprised even you?
A: That started as a joke, and then it turned out to work! If you tell people that this TV set shows only news, and that one only entertainment, and a third is a normal all-purpose TV, then even when you show the same programs on the TVs, people will judge the news on the all-news set smarter, and the comedies on the all-entertainment set funnier, than those programs when seen on “regular” TV! The same programs. Participants even judged the picture quality to be better on the “specialist” sets.
Q: That’s the halo effect you write about.
A: Right. When we think something about a person is better than other people, we start assuming everything about him is better. If your TV set is smarter or funnier than most, naturally it’s better looking too.
Q: The human-computer interaction convinced you our emotions are pretty simple.
A: We think we need expertise and therapists to really grasp the complexity of human emotion, but it turns out there’s really only two questions you have to ask yourself about someone’s emotion. Are they happy or sad? Are they aroused or calm? We don’t have to do an incredibly complex calculation of exactly what nuance of what emotion they’re feeling. All we have to know is those two things and act accordingly.
Q: What does that entail?
A: When we put people in front of computers with driving simulators, with other drivers cutting them off, we found misery loves miserable company. If you want to cheer up an unhappy but calm person, be very slightly happier than him, then slowly shift your emotions to be happier and happier and that person—because we mimic the emotions of those around us—will shift with you.
Q: What if they’re actively unhappy?
A: We found that when the computer urged people to vent after they were cut off, saying, “Let them know how you feel, get it out of your system,” or had the computer vent for them, so that the car would scream, “You idiot! Get off the road!”—it actually triggered more negative experience. Venting activates the negative part of the brain, sends it into higher gear. A much better strategy is to show that you’re really listening, which causes people to move on into the problem-solving and rational part of their brains, which actually calms them down and makes them feel better. When bad drivers made the subjects swerve, the car would say, “Gee, that was great driving,” or even, “That car is badly designed; in fact he couldn’t see you,” and that took people out of the search for someone to be mad at.
Q: Do you think that finding has implications for workplaces?
A: When you deal with an angry employee you want to be much more thoughtful than letting them rant. You need to get them thinking what can be done, rather than whose fault is it. Focus on the details as opposed to the generic feeling, and start to lever people out of the angry part of the brain.
Q: How should people respond to the so-called critic phenomenon: we like those who praise but we think critics are smarter?
A: Studies of film and book reviews show that readers rate negative critics as more intelligent, even if they don’t like them. Similarly they don’t like people who praise themselves, but they take modesty at face value: when you don’t toot your own horn people assume you have good reason. The same situation goes for the office, and it often seems you can be liked (when you praise others) or respected (for criticizing others), but not both. The best strategy for countering this is a mutual admiration society. I’ll say nice things about you to everyone, and you’ll say nice things about me, and we’ll both get the benefits of praise, and the benefits of praising.
Q: You believe negativity is underestimated in standard management techniques like the evaluation sandwich: praise, criticism, praise.
A: When we hear something negative we immediately and literally forget what happened just before, and we remember much more what happened after. If I start out as a manager with effusive praise, which supposedly makes the coming criticism go down easier, the minute I move on to the negative things that praise is blasted from your brain, doesn’t make it into memory at all. You don’t remember it, and what are you hearing? Negativity. Even if you then hear some vague—“Basically you’re doing fine”—afterwards, what you remember is vicious criticism, because criticism always seems much worse than it is, and you are very unhappy. And I, as manager, am thinking, “Wait a minute, I just praised this guy. Didn’t he hear me?” And the answer is, he heard it, but his brain was built to forget it.
Q: Is that why we keep doing the same dumb things over and over; “once burned, twice shy,” isn’t true?
A: It’s more the opposite. We remember the negative experience far more than the cause, if it comes right before. And then, when we’ve burned our fingers again, we say, “Damn, now I remember, that happened before!”