"It is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility." - Hans Moravec, AI researcher
Humanoid craze overblown!
Rodney Brooks, the Roomba co-creator and a Godfather of Robotics, has warned that the current rush for human-like robots is fueled by hype. He predicts that in 15 years, massive amounts of money will have disappeared trying to force performance from these machines. He believes that general-purpose humanoids are not coming home soon.
Form over function
The issue is that a human shape makes a promise the robot cannot keep. When people see a humanoid form, they expect it to act like a person. Current robots perform specific tasks well but fail at general human capabilities. This mismatch leads to inevitable disappointment when the machine cannot do everything a human can.
Click to see our collection of posts on Robotics
There are limits
Brooks stated he would not stand near a humanoid robot. If these heavy machines lose balance, their powerful mechanics make them dangerous to anyone nearby. He argues they are not robust enough yet to be safe in our homes and are prone to falling over unpredictably.
A problem of skills
A major hurdle is replicating human touch. Humans lack the data language to teach robots how to feel and manipulate objects. Visual data alone is insufficient to recreate the nuance of human finger movement found in nature. Humans use complex feedback loops of touch that robots simply do not have yet.
So what will the future be like
Instead of sci-fi androids, the future holds specialized robots on wheels. They will work alongside us in warehouses, helping with specific tasks rather than replacing us. Brooks expects practical tools rather than artificial people, suggesting we will see helpful machines with multiple arms rather than walking humanoids.
Summary
Godfather of Robotics Rodney Brooks argues that general-purpose humanoid robots are overhyped and years away. He highlights safety risks, the lack of touch-based data for dexterity, and the disconnect between appearance and capability. He foresees specialized machines rather than human replicas.
Food for thought
If we cannot teach robots the language of touch, can they ever truly understand the physical world we inhabit?
Click to see our collection of posts on Robotics
AI concept to learn: Moravec's paradox
Moravec's paradox is the discovery that high-level reasoning requires very little computation, while low-level sensorimotor skills like walking require enormous resources. This explains why we have AI chess grandmasters but struggle to build robots with the dexterity of a toddler.
COMMENTS