What if our next generation of social robots could help people regulate their emotions—not by mimicking humans, but by behaving more like horses?
Key Points at a Glance
- Researchers at the University of Bristol draw inspiration from therapy horses to design emotionally responsive robots.
- Unlike current companion robots, these would resist interaction unless the user is emotionally regulated.
- The study challenges the idea that social robots should always be compliant and comforting.
- This approach could revolutionize emotional therapy and human-robot interaction across many fields.
In a quiet revolution of robotics design, researchers from the University of Bristol are taking notes from an unexpected teacher: the horse. Their latest study suggests that the emotionally reflective behavior of therapy horses could inspire a radical new generation of social robots—ones that are not just comforting, but transformative.
The research, led by Ellen Weir and presented at the prestigious CHI 2025 conference in Yokohama, draws from Equine-Assisted Interventions (EAIs). These therapies are often used to help individuals with PTSD, trauma, and autism—especially those who struggle to communicate and regulate emotions verbally. EAIs work because horses, unlike conventional therapy tools, respond to emotional energy and body language. They won’t follow commands from someone who is emotionally unsettled, but once the individual becomes calm and grounded, the horse engages willingly. This mirrored feedback helps patients gain insight into their emotional states and learn regulation skills.
Weir and her colleagues believe that therapeutic robots should adopt a similar model. Today’s social robots typically follow a script: they are compliant, friendly, and endlessly patient. But what if they were more autonomous—more discerning? What if they refused to engage until their human partner reached a calm emotional state?
“We found that therapeutic robots should not be passive companions but active co-workers, like EAI horses,” said Weir. “They could help users become more self-aware by resisting engagement until the user is emotionally regulated.”
This concept could transform not only emotional therapy, but also applications like social skills training, emotional coaching, or even corporate wellness programs. By creating interactions that depend on emotional reciprocity, these robots would no longer just mimic empathy—they would help cultivate it.
Of course, this bold new direction comes with complex challenges. Can robots truly read and respond to human emotional states? Weir believes it’s possible—but only with serious advances in emotional sensing technologies, motion design, and machine learning. “The next challenge is designing robots that can interpret human emotions and respond dynamically—just as horses do,” she said.
And there are ethical questions, too. Can a machine ever replicate the emotional depth of an animal? If it can, how do we ensure these interactions remain meaningful and not manipulative? The team acknowledges that replacing sentient animals with machines opens a Pandora’s box of ethical concerns—but insists that, with the right design principles, therapeutic robots could offer an accessible alternative for those unable to benefit from traditional EAIs.
By flipping the design script—shifting robots from obedient assistants to emotionally intelligent mirrors—Bristol’s researchers aren’t just reimagining therapy. They’re reimagining connection itself.
Source: University of Bristol
