LinkedIn user Nick Dimichino recently witnessed some disturbing behavior on a flight. He observed another passenger using Claude, the chatbot developed by the San Francisco-based company Anthropic, and verbally abusing the chatbot for three hours.
Would we want to hire or associate with someone we knew beat, rape, or abuse robots? We certainly wouldn’t want to hire or associate with someone who tortures cats or beats their children or partner. We would probably call the police. But what about robots or AI language models?
Children tend to be quite rough with voice assistants like Siri or Alexa. Since these assistants always respond in a friendly and helpful manner, children test their limits by hurling increasingly harsh insults at them. Good parents, however, want their children to say “please” and “thank you.” But do you really have to do that with an AI or a robot? Especially when we hear from companies that these polite phrases consume processing time—and thus energy and cooling water. Environmentally conscious parents who also value politeness find themselves in a dilemma here.
Japanese robotics researchers observed children in a shopping mall blocking and mistreating a service robot by standing in its way and even starting to hit it. The Japanese researchers’ solution to prevent this was to program the robot to seek out the presence of adults as soon as small children approached it. This is because adults tend to deter children from acting aggressively toward the robot.
Patrick Lin, director of the Ethics and Emerging Sciences Group at California Polytechnic State University, sees the use of robots for these purposes as a potential step backward. This is because it makes us humans less human, as we resort to violence instead of negotiating and reaching agreements.
In a TED Talk, MIT robotics ethicist Kate Darling describes how she showed a friend a small dinosaur robot. This little robot, named Pleo, had several sensors and motors built in; it could walk around and move its head, but it could also tell whether it was standing upright or lying down. When it was lying down or hanging upside down, it started to cry. Darling’s friend examined the robot and held it upside down, which caused the robot to cry. Darling felt so uncomfortable that she took the robot back from her friend. Her reaction—this compassion for the dinosaur robot, which was ultimately nothing more than a toy—surprised even her, and she asked herself why we form emotional connections with machines.
We can see another example in the following video, where a four-wheeled delivery robot from Serve Robotics—lacking hands—is unable to press the button on a traffic light and therefore asks bystanders for help. One of the passersby refuses to help the robot because he believes that such robots take away people’s jobs and their place in the world. He doesn’t stop there; he even insults the robot and tells it what he thinks of it.
If one person were to insult another in this way, we would certainly take notice and perhaps intervene. But we would be unlikely to side with the aggressor. If he were to start hitting the other person for no reason, that would be the moment for us to call the police.
Since people tend to anthropomorphize machines—whether in the form of robots or chatbots—it’s understandable why our reaction to this abuse, as seen in the examples, triggers such visceral responses in us. We empathize with them and condemn those who inflict this suffering.
As a result, we will likely also have to discuss the extent to which robots or AI models should be granted rights in the future, so that such abuse cannot go unpunished and, as a result, is not directed at humans as well. Someone who abuses a machine is likely to stop at nothing when it comes to other people.