The first Humanoid with ChatGPT integrated.

Further advances in AI-machine integration.

Pepper is a famous humanoid robot that received an enhanced brain with generative capabilities and was released into the real world. The experience provided a unique window into understanding how humans react when science fiction collides with everyday life.

Researchers from the University of Cambridge took Pepper to a crowded festival, equipped with microphones, voice recognition, the power of ChatGPT, and voice synthesis. The robot became, in practice, a walking chatbot ready to converse with anyone.

Eighty-eight people were interviewed, each interacting, asking questions, testing, and above all judging: Could a robot like this fit into our real lives?

The complexity of teaching a robot that it is human.

The result was a roller coaster of emotions. Some were amazed by Pepper's gestures, eye contact, and politeness, but there was also frustration when the robot talked too much, interrupted, or failed to understand a more subtle question.

Some found its humanoid form strange, contrasting with basic limitations such as not recognizing facial expressions or reacting to a simple touch on the shoulder. Pepper gestures and moves its head, but it still does not interpret human body language as we expect. However, the most impressive thing was the result of the study, which revealed a key point.

The more a robot resembles a human being, the more we expect it to behave like one, even knowing how to wait its turn to speak, understand jokes, react to humor, or even know when to remain silent. When that fails, discomfort arises, a breach of expectations that generates that strange Black Mirror-style feeling.

Although it still struggles with naturalness, the experience was valuable. Social robots like Pepper could appear in hospitals, stores, and even schools, and knowing how first impressions shape our acceptance is vital so that they are not just cute dolls, but true companions. In the end, Pepper showed that putting generative AI in a humanoid body is only half the battle; the other half is teaching these machines to understand our invisible rules of coexistence, the same ones that maintain a good conversation between two humans.