HUMANOID ROBOTS
MYTHS, FEARS AND REALITY
This text was inspired by two recent events: the presentation of humanoid robots from TESLA, and the publication of Patric M. Rael's book “UTOPIA ANDROIDIA” about a potential society integrating people and robots.
The presentation mentioned above is obviously intended to demonstrate the progress in the development of humanoid robots, the most notable events of which were the ASIMO robot (Honda Corporation, 2000) and ATLAS (Boston Dynamics, 2013). The past quarter century has demonstrated progress in what the humanoid robot looks like and how it moves - apparently, it is increasingly similar to how people look and move. However, what the public has yet to see is the use of these developments in practice.
The respectable public seems to be under the impression that the reason these robots were not used "for their intended purpose" is that they do not look human-like enough or do not move dexterously enough - and now these problems have finally been solved, and millions of mass-produced robots are about to significantly change the way of life of humanity. And such radical changes cause natural fears - which was the reason for writing the above-mentioned "UTOPIA ANDROIDIA", which suggests ways to overcome potential problems.
But is the lack of human-likeness really the reason why such robots are not used yet? As is often the case, it is helpful to look at what the presentations and descriptions are silent about to answer this question. What they are silent about is precisely those aspects that determine where, how, and why robots are used - or not - namely the utilitarian capabilities of robots and their cost-effectiveness.
Humanoid robots are not the only possible and existing ones! Half a century ago, industrial robots began to be used in the production of cars, replacing people in physically difficult operations of welding body parts. Then, robots replaced people in the assembly of electronic products. Today, robots are mastering container handling operations in ports. What is common in all cases is that, while replacing people, these robots do not copy the "design" of a human. Still, their design and capabilities are specialized for their area of application. Why is this so? Because it is reasonable: abandoning a humanoid form in favor of a specialized one increases productivity, and abandoning versatility reduces the cost of robots. Having looked at your household, the reader will find that robots already exist - in the form of washing machines and dishwashers, automatic vacuum cleaners, automatic solar energy control machines, cruise control in a car, etc. What can a humanoid robot add to this? Obviously, it cannot replace either a washing machine or a dishwasher; replacing the TESLA or Mercedes automatic driving system with an android does not add new capabilities, but reduces the number of seats in the car by one and doubles the total cost.
These examples clearly point to two key factors that determine the scope of the android's usefulness: the expansion of capabilities and the cost of this. The basic idea of an android is not to replace home appliances but to complement their capabilities, playing the role of a servant. To do this, it must not (optionally) look like a human but also be able to do what is self-evident for a person:
operate safely in a human environment, i.e., move among disorderly furniture, scattered toys, and running children
be trained by demonstrating the actions that it must perform at the request of the owners
be able to distinguish between reasonable commands and, for example, demands from children to give them a couple of pounds of candy and set off fireworks in the bedroom
Of course, demonstrated androids do not have these capabilities (otherwise, they would have been demonstrated); moreover, there are yet to be well-founded ideas on how to implement these capabilities. But even if we assume this will become feasible shortly, the question remains: why should we use an android to carry dishes from the table to the dishwasher if we then go to the gym to ensure a healthy minimum of physical activity?
That is, there is a real lack of demand not only for those beautiful but primitive in capabilities androids but also for future more advanced ones.
If we are talking about industrial robots, copying of universal but very limited physical capabilities of a person by androids means that they can be used only as an addition to specialized equipment; the android moves levers and presses buttons, controlling some equipment, using its actuators. Signals from the android "brain" control the android's body. The body of the android manipulates the controls of the equipment, which convert the movements of the levers and knobs into electrical signals that control the equipment. It does not take much intelligence to figure out: it is possible to connect the "brain" of the android directly to the equipment, excluding both the body of the android and the controls of the equipment. Which is actually happening in the real, not mythical world.
Thus, the future of androids looks bleak: there is no real need for them, and they are innately ineffective economically. Fantasies about androids are born of our imagination, like the mythical Hercules and Icarus or time travel; since androids are, in principle, feasible, there are attempts at implementation, but there are no objective grounds for fear of their widespread use and radical change in society. The consequences of using artificial intelligence in other ways will change society no more radically than the disappearance of telephone operators and telegram carriers - if the respectable public still remembers what we are talking about.
Why, then, are significant resources spent on developing androids? Because myths often defeat rationality, there are those who are ready to finance myths. If you have money to develop an android, there will be those who will undertake it and those who will predict the success of the idea.