A study from Imperial College London shows that people feel sorry for Artificial Intelligence (AI) bots and try to include them in games if they see the bots being left out.

The Imperial College researchers used a game called “Cyberball” where players throw a virtual ball to each other. They had 244 people play this game with an AI bot involved.

The researchers describe the methods and results of the study in a paper published in Human Behavior and Emerging Technologies. When the AI bot was excluded, the human players often tried to throw the ball to the bot more to make up for it.

This behavior is similar to what people do when they see other persons being left out: they try to include them more.

This reaction shows that humans can treat AI like social beings. Even though the participants knew they were playing with a bot, they still wanted to be fair to it.

Social AI design

This tendency to sympathize with AI could be important for designing future AI systems, suggesting that as AI becomes more common in daily life, people might interact with them as they would with other humans.

As AI technology grows, and bots or virtual agents become more involved in human activities (like work or social interactions), understanding this human behavior can help in making these interactions smoother and more natural. If humans naturally want to include and be fair to AI, this could lead to better collaboration between humans and machines.

“This is a unique insight into how humans interact with AI, with exciting implications for their design and our psychology,” says researcher in a press release issued by Imperial College.

The researchers believe that we should think about how we design AI to not just perform tasks but also to engage socially in a way that feels right to humans. They’re looking into more ways to study this, like having face-to-face interactions with AI in different settings.

Let us know your thoughts! Sign up for a Mindplex account now, join our Telegram, or follow us on Twitter.