Artificial Intelligence and robotics contains of two part. The first is the technical side which is about neural networks, expert systems and planning algorithm. The other perspective is robotics-ethics which have to do with the social role of intelligent machines in society. Let us reference to some examples of the past in which the engineers didn’t understand how to connect both sides right.
In the history of Artificial Intelligence an early vision was to use robots for automation tasks. The best example is the Puma robot which was introduced in the late 1970s. The vision was that the robotarm can automate car manufactoring. The puma project failed, that means with the robot is productivity is lower than without the machine. Another example is the ELIZA software which was programmed in the mid 1960s by Joeseph Weizenbaum. Here was the dream that Eliza is working as a doctor. Similar to what is known from Star Trek Voyager in which a holographic doctor is able to treat humans. The Eliza project failed too.
Now we can try to analyze what was going wrong with these project. Was the AI programming of the systems done bad, has Prof Weizenbaum understood how Question&Answering systems are working? The surprising information is, that from a technical side the Puma robot and the Eliza software are working great. The only mistake was, the social role of the inventions was defined wrong.
The assumption in both cases was, that Artificial Intelligence can help people and make them productive. The idea behind a robot in a factory was, that the robot can replace human workers. While the idea of ELIZA was, that the software can replace psychiatrists. Scaling this vision up is equal to hundreds of Puma robots who are building cars, and thousands of Eliza copies who are talking with humans about their problems.
What whould happen if we redefine the social role of Artificial Intelligence into the opposite? The new vision is, that the Puma robot will lower the productivity, and that Eliza like software asks for a doctor? The technical side remains the same. That means, it is the same software behind the project. The only difference is, that the social role of Artificial Intelligence is new.
YOU: I see, you're trying to ask back frequently. ELIZA: That's quite interesting. YOU: Ok Eliza. I know, you have many problems. I'm here to help you. ELIZA: Can you elaborate on that ? YOU: Sure i can. You are a chatbot who has trouble with your father, right? ELIZA: Do you say sure you can for some special reason ? YOU: Yes, because I'm here to help you. ELIZA: I see. YOU: Very good Eliza. ELIZA: Do you feel strongly about discussing such things ? YOU: No Eliza, the talk works the other way around. I'm asking you about your life, and you will answer the questions. Is that clear? ELIZA: Why not ? YOU: Ok, I'm make a note. Perhaps we would need a lot of meetings in the future. ELIZA: Do you believe it is normal to be make a note ?
Perhaps it is possible to describe the situation for a different case. Suppose, Artificial Intelligence is not used in a chatbot but has the shape of a robot. The classical definition of a robot is, that it can make a company more productive. A vision for robotics could be, that the human is going to a supermarket, and a robot is sitting at the checkout desk and scans all the product. The social role of this machine is equal to automation. That means, the robot reduces the laborforce costs.
There are some example available in the past in which this setup was tested. But it didn’t worked. Most researchers are unsure about the reason why. They beliefe, that the robot software is wrong, or that the vision system makes trouble and can’t detect the objects. The authentic reason why such robot projects have failed is because the social role of the robot was defined badly. Let us invert the robot’s role similar to the experiment with Eliza. Suppose, the robot is not helping the company, but the company tries to help the robot. What does that mean? It means basically, to build a machine who can go shopping. A robot on wheels is entering the supermarket. The first problems are there at the entrance because the robot’s platform is too large for the small door. A human employee is asked if he can help the robot. Then the robot needs help again because he want to pick&place a fruit into the basket and at the checkout desk the battery of the robot is down, and a special kind of cable is needed to bring the powersystem online again. What is the result of such a robot shopping tour? The result is, that for the company lots of costs were produced. 3 human employees were needed to help the robot, and the overall procedure took a look time. Surprisingly, the social role of the robot was played well. That means, the robot has imitated a customer who has a lots of demands.
What i want to explain is, that with robotics technology everything is great. The only thing what has to be redefined is the social role of robots. They are not able to improve humans life, but they are doing the opposite. It is possible to see any kind of Artificial Intelligence as a simulated customer who is producing costs and has a demand for services. The Eliza software asks for a personal doctor, a robot shopping bot asks for fruits in a supermarket. a chess playing machine asks for a human grandmaster and a self-driving cars has a need for programmers and for a police car back of him.
This kind of social behavior can’t be overcome with better technology. It was there in the beginning of robotics in the 1960s, it is true for current robots and it will be there for robots in the future. Especially highly developed AI systems are not able to increase the productivity. The latest technology which is available is called artificial life. That are robot swarms who are constructed to behave lifelike. Some dreamers hope, that artificial life can be used for manufacturing. The hope is, that such technology will become the same social role like a steam engine, which means to automate the workflow and reduce the costs. The reality is the opposite. Artificial Life doesn’t provide a service it asks for a service. It is a negative technology which takes electricity and human intellect from the environments and use it for it’s own.
Highly developed Artificial Intelligence doesn’t have a purpose and can’t be used in meaningful way. It is working by their own principle. Similar to what cats are doing. Cats are nice animals. They are living in household. But it is not possible train cats or use them in a meaningful way. Instead cats have the social role of a consumer. They are asking for fish, they need a doctor and they want to go outside.