Scientists Build a Rat-Like Robot That Can Interact With Real Rats and Influence Their Emotional States

Robots are becoming increasingly advanced, but making them interact naturally with living organisms remains a major scientific challenge. While modern robots can recognize objects, track movement and respond to their surroundings, reproducing the subtle signals that animals use during social interactions is much more difficult. Now, researchers have developed an autonomous, rat-like robot designed to interact with real rats by imitating their physical structure, movements and social behaviours.

The research, led by Guanglu Jia and his team, introduces a new approach to studying animal behaviour using biohybrid systems. These systems combine living organisms with artificial technologies, allowing scientists to investigate biological processes in a more controlled and measurable environment.

The rat-like robot can attract the attention of real rats, interact with them for extended periods and even influence their emotional states through different interaction patterns. The findings could help researchers better understand how social behaviour works and how artificial systems might eventually interact with humans more naturally.

Why Are Scientists Building Robots That Behave Like Animals?

Social behaviour plays an important role in the lives of many animals, including rats. They communicate through body movements, physical contact, changes in posture and other behavioural signals. These interactions can influence how animals respond to one another and how they experience their surroundings.

However, studying these behaviours under natural conditions can be difficult. Living animals do not always respond in predictable ways, and their interactions can change depending on their environment, emotional state and previous experiences.

Traditional experiments often allow researchers to observe animal behaviour, but controlling every aspect of a social interaction is challenging. Scientists may struggle to determine precisely which movement or signal caused a particular response.

Robots could help address this problem.

Unlike living animals, robots can be programmed to repeat specific movements and interaction patterns under controlled conditions. If a robot can communicate with an animal in a sufficiently natural way, researchers can systematically change its behaviour and observe how the animal responds.

This approach could provide valuable insights into the biological mechanisms responsible for social interaction, attraction, emotional responses and communication.

How Does the Rat-Like Robot Work?

The research team designed the robot by studying the anatomy, movements and social interactions of real rats. Instead of creating a conventional machine that simply resembles a rat, the researchers aimed to develop a robot capable of behaving in ways that would interest and engage living rats.

The system combines several important technologies, including imitation learning, visual perception, target tracking and autonomous behavioural decision-making.

1. Learning Social Behaviour From Real Rats

One of the most important features of the robot is its use of imitation learning.

Imitation learning is a technique in which a system learns how to perform certain actions by observing demonstrations. Rather than relying entirely on manually programmed instructions, the robot can learn behavioural patterns from examples of how real rats move and interact.

By studying animal demonstrations, the researchers developed templates that helped the robot reproduce subtle patterns of social behaviour.

These patterns are important because animals do not respond only to physical appearance. The timing, direction and nature of movements can also influence whether another animal pays attention or approaches.

A robot that looks like a rat but moves unnaturally might fail to attract much interest. By incorporating behavioural information from real rats, the researchers aimed to make the robot a more convincing social partner.

The resulting system could capture the attention of living rats and significantly increase their interest in interacting with it.

2. Seeing and Tracking Real Rats

Another major component is the robot's visual perception system.

To interact effectively with a freely moving animal, a robot must be able to observe its surroundings and identify where the animal is located. It also needs to track the animal as it moves, rather than relying on a fixed position.

The researchers integrated visual perception and target tracking to help the robot follow the behaviour and movements of real rats.

This capability allows the robot to respond to changing situations during an interaction. Instead of repeating the same action regardless of what the rat does, it can use visual information to guide its behaviour.

This is an important step toward creating autonomous robots that can interact with living organisms without requiring constant human control.

3. Making Decisions During Social Interactions

Recognizing and tracking an animal is only part of the challenge. The robot must also decide how to behave in response.

The system therefore combines its perception capabilities with behavioural decision-making. This allows it to select actions based on the ongoing interaction instead of following a completely fixed sequence.

Such responsiveness is particularly important when working with animals because their behaviour can change rapidly.

By combining learned social behaviour, visual tracking and autonomous decisions, the researchers created a robot capable of participating in more dynamic and sustained interactions.

A Robot That Can Interact With Rats for 30 Minutes

One of the study's notable achievements is the robot's ability to interact continuously with freely behaving rats for approximately half an hour.

Maintaining an interaction for this length of time is significant because brief encounters may not provide enough information to understand how animals respond to artificial social partners over time.

The robot's ability to sustain repeated interactions gives researchers more opportunities to observe behavioural changes and investigate how different actions affect the animal's responses.

It also addresses a limitation of conventional social experiments, in which researchers may have difficulty reproducing exactly the same interaction conditions across multiple trials.

Because the robot can repeat its behaviour, scientists can examine how rats react to particular movements or interaction patterns under comparable conditions.

This repeatability could make future experiments more systematic and help researchers distinguish between different factors that influence animal behaviour.

However, the ability to maintain an interaction does not necessarily mean that a rat perceives the robot in the same way it perceives another living rat. Understanding the differences between responses to biological and artificial partners remains an important question for future research.

Can a Robot Influence an Animal's Emotional State?

Perhaps the most intriguing finding is that the researchers demonstrated the robot's ability to modulate the emotional states of rats through different interaction patterns.

In other words, the way the robot interacted with the animals could influence their internal states.

This finding is important because social behaviour involves more than simply moving toward or away from another individual. Interactions can also affect how an animal responds emotionally to its environment.

By changing the robot's behaviour and observing the rats' responses, researchers can investigate how particular social signals may be associated with changes in emotional state.

This creates an opportunity to study the relationship between external behaviour and internal biological processes in a more controlled setting.

Nevertheless, influencing an animal's emotional state does not mean that the robot understands the animal's feelings in the human sense. The system uses behavioural patterns and automated decision-making, while scientists interpret the resulting biological responses.

Further research will be needed to determine precisely which robot behaviours produce particular effects, how long those effects last and whether the findings apply to different animals or social situations.

What Does This Mean for the Future of Artificial Intelligence?

The implications of this research extend beyond animal experiments.

Artificial intelligence systems are increasingly being designed to interact with people through conversation, digital assistants, social robots and other interactive technologies. Yet natural communication involves much more than exchanging words or processing commands.

Humans respond to timing, gestures, facial expressions, body language and the behaviour of others. These signals can influence attention, trust, comfort and emotional responses.

The rat-like robot offers a possible framework for investigating how artificial systems can influence living organisms through carefully designed interactions.

By studying these processes in animals, scientists may gain a better understanding of the principles that make interactions between biological and artificial systems effective.

In the future, similar approaches could potentially contribute to the development of social robots for healthcare, assisted living, education or other environments where natural interaction is important.

However, translating findings from rats to humans will require extensive additional research. Human social behaviour is considerably more complex, and the results of animal experiments cannot automatically be applied to human relationships.

There are also ethical questions to consider. If artificial systems become capable of influencing emotional states, researchers and developers will need to examine how such capabilities should be used responsibly.

A New Tool for Understanding Social Intelligence

The development of the rat-like robot represents an important step toward integrating robotics with the study of living behaviour. By learning from real rats, tracking their movements and making autonomous decisions, the system can participate in sustained social interactions and influence aspects of the animals' internal states.

Its most significant contribution may be the experimental opportunities it creates. Rather than merely observing animals interacting with one another, scientists can use a programmable artificial partner to investigate how specific behaviours affect biological responses.

Although much remains to be learned, this approach could help researchers uncover the mechanisms behind social behaviour and develop a clearer understanding of how living organisms respond to artificial partners.

Ultimately, the research raises a fascinating question: if a robot can learn to interact with an animal in ways that influence its behaviour and emotional state, could similar principles help us understand how humans will respond to increasingly social forms of artificial intelligence?

The answer remains uncertain, but this research offers a new way to investigate the relationship between biological intelligence and machines.

Reference: Jia, G., Chen, Z., Zhang, Y. et al. Modulating emotional states of rats through a rat-like robot with learned interaction patterns. Nat Mach Intell 6, 1580–1593 (2024). https://doi.org/10.1038/s42256-024-00939-y