“Fake connections”: Pass laws around public interaction with AI, say KU Leuven scientists

Patients should always have the choice to be assisted by a human being and not by AI, according to a new position paper published by a group of KU Leuven researchers. The group is alarmed by the pace of “robotisation,” they say, particularly when it comes to medical care.

“We are living in a very big social experiment,” the authors write. “More and more, we have to interact with digital and physical robots instead of people. Without having chosen it or even having thoroughly considered the implications.”

The position paper was authored by nine scientists and engineers, including famed psychologist Dirk De Wacther and Yvonne Denier, associate professor of healthcare ethics and co-director of the Ethics@KU Leuven community . 

To address the “many ethical issues of artificial intelligence and robotics,” the authors have a concrete request: Incorporate the right to human contact in national and international law. “Then it is legally enforcable,” they write.

This should be applied much broader than a medical context, according to the authors. It should also be enforced in workplaces, for instance, in order to prevent non-human relationships between employee and employer. Human contact is a fundamental need, they say, “for our physical and emotional wellbeing”.

 De Wachter recently took part in the investigative news programme Pano to see how an AI version of himself would treat patients. He was shocked by how realistic it was and pointed to a “fake connection between patient and therapist.” You cannot simply replace “the caring interaction with a fellow human being,” he said.

In a broader sense, “there are many well-known ‘system failures’, like algorithms that wrongly assess the risk of recidivism (reoffending) as higher for black people. Or facial analysis that works better with some groups of people than others. Technology is always developed in a specific political and economic context.”

The authors admit that many tasks can be done more easily and quickly by AI, but issue a warning in terms of people abandoning learning to think strategically and critically, which is part of normal human development, they emphasise.

Any kind of customer, client or patient should always have the choice to be assisted by a real person, the authors say. Pointing to companies using AI as a cost-saving device, the authors suggest a “robot tax” to encourage continuing to employ people to work with the public and with each other.

But they also want to make it clear that they are not against the development of robotic or AI tools, which can revolutionise health care and other fields. “Ultimately it’s about asking ourselves what kind of society we want to live in,” they write. “Do we want to risk losing our humanity in the ambition for more efficient technologies, or do we want to live in a society that allows space for the complexity, unpredictability and authenticity that comes with human interaction?”

Read the position paper here (in Dutch)