Experts Warn of Risks as Humanoid Robots Advance Toward Reality

The development of autonomous humanoid robots is progressing, but experts caution that without stringent regulations, their potential may pose more risks than benefits. Dr. Carl Strathearn, a lecturer in Computer Science at Edinburgh Napier University and an advisor on AI and robotics to the UK Government’s Office for Science, will share these insights at the upcoming New Scientist Live event, scheduled for October 18-20 at London ExCel.

Despite impressive demonstrations of humanoid robots performing tasks such as pouring drinks and mimicking facial expressions, Dr. Strathearn emphasizes a significant gap between these showcases and the reality of robots that can reliably assist in everyday life. He states, “The biggest problem is the lack of real-world data and the technological means of gathering it in large enough quantities to train our robots effectively.” Presently, many robotic systems depend on virtual simulations, reinforcement learning, or even analyzing videos of human activities, leading to machines that perform adequately in controlled environments but struggle in the complexities of daily life.

During the event, Dr. Strathearn plans to present his robotic creation, Euclid, which he describes as a “friendly robot.” He illustrates the challenges by using a simple object like a cup, noting that variations in size, weight, shape, and color present a daunting task when considering the multitude of objects in a household.

One proposed solution is to crowdsource real-world data on a large scale, potentially via devices like Meta’s Ray-Ban smart glasses. However, Dr. Strathearn acknowledges the ethical complexities and logistical challenges of gathering data from thousands or even millions of individuals.

While science fiction often depicts robots as threats, Dr. Strathearn argues that the real danger stems from human misuse of these technologies. “If you are talking Terminator, the answer is no. We are and always have been more of a danger to ourselves than anything else,” he explains. As he leads a petition calling for regulations regarding humanoid robots in public spaces, he highlights increasing incidents of near misses involving humans and robots, attributing these issues to human behavior rather than AI flaws.

The issue of perception also poses a challenge. Robots that closely resemble humans risk entering the “uncanny valley,” which can elicit discomfort. Yet, in specific scenarios, such as dementia care, a familiar human-like appearance may provide comfort. Dr. Strathearn notes, “People have different thresholds of perception when it comes to creepiness,” leading to a diverse range of robot designs, from highly lifelike to those with minimal features.

His research during his PhD led to the creation of the “Multimodal Turing Test,” which examined whether communication through lifelike robots influences perceptions of AI as more human. Subsequent studies in Japan confirmed that individuals were more inclined to view AI as human when interacting with realistic robots. Dr. Strathearn believes that fostering acceptance of humanoid robots will require careful introduction and education, particularly for children learning about robotics and AI in schools.

Despite the challenges highlighted, companies are rapidly advancing in the field. Dr. Strathearn cautions, “The hype is a major issue. We are far from humanoid robots that are good enough to do everyday tasks effectively, but that doesn’t stop major companies wanting to mass produce them.” He notes a critical skills shortage in robotics, exacerbated by universities often segregating students into specific disciplines like computer science, engineering, and design. “Without a solid foundation in education, I worry about the sustainability of the humanoid robotics industry,” he adds.

Interestingly, Dr. Strathearn identifies space exploration as a potential area where humanoid robots may prove beneficial sooner than expected. He remarks, “Space exploration for sure — we could use telemetric or AI-controlled humanoids to work in space for longer periods than humans, advancing us further into the unknown.” Future humanoids could even assist in terraforming planets or exploring challenging terrains beyond the capabilities of existing robotic rovers.

In conclusion, while the potential for robots to aid in colonizing new worlds exists, Dr. Strathearn stresses that ensuring their safety and reliability through strict regulation is imperative before they become commonplace in society. He asserts, “Robots might terraform Mars one day. But on Earth, only strict regulation will keep us safe.”