People those are interested in science fiction has recently received a true double treat when the latest projects of Boston Dynamics and Google have been launched. New moral problems have come to the rescue!
We do not know if it was the first time that pop culture began to deal with this issue, but in 1982, when Blade Runner came out, many were concerned about the following problems:
Who are the robots? Who are human beings? Where are the boundaries between them and how can we distinguish them?
In those days, such questions were superficial and philosophical. At that time, the first Macintosh was not released.
When you look at it, a picture will come to mind that the boundary between a person and a computer is very clearly defined. Now, 35 years later, new branches of science have emerged, including Google’s latest invention in the field of artificial intelligence that can halve the human body with a single phone.
The above video was presented by this company at this year’s Google I / O that took place less than a week ago. The concept was such that nothing was pre-arranged or mounted. So, the employee of a hairdresser regularly picked up the phone and talked with artificial intelligence all the time, she thought to talk to an assistant trying to schedule an appointment for her boss.
It should be mentioned that this is a really simple and uniform conversation with pretty predictable and template questions and answers. But, in any case, it’s amazing that Google’s assistant can go through slower parts of the conversation. In the second shot you can hear Duplex reserving a table in the restaurant. The worker who received the call was overwhelmed with the requirements.
As you can hear, Duplex can not only deal with the situation, but could also understand a complex problem such as that the reservation is only possible for more than five people. He also asked about the waiting time and finished the conversation. The big step forward for this technology is precisely that the machine can lead an active conversation.
Now we can almost certainly say that distinguishing people from machines could soon become a daily problem.
Voight-Kampff Test
The aforementioned “Blade Runner”, directed by Ridley Scott, starts with a glorious scene in which one of the exterminators performs the Voight-Kampff test on a replicant, i.e. Android, as it turned out later. The main goal of the test, which was designed for psychological testing, was to distinguish people and robots, since robots became so developed that it was impossible to distinguish them using conventional methods.
Of course, in reality there is no Voight-Kampff test. It is just a product of fiction from the pen of the famous sci-fi writer Phillip K. Dick, who actually wrote the book “Are They Dreaming of Android’s Electric Sheep?”, Which later served as the basis for this film. This fictitious test is based on psychology. The International Affective Picture System is a photography base that is designed to provide a standard set of photos for studying emotions and attentions. For disturbing photos that can be found in the system, many people have emotional reactions. Robots, of course, do not have this possibility, which confirms the assumption that they are followed by the absence of real emotions.
Although the Voight-Kampff test is not realistic, Turing test definitely is. It was developed by Alan Turning, a British mathematician and father of modern information science. The test serves for the “assessment” of artificial intelligence. If one who’s estimate cannot reliably distinguish a machine from a man, it means that he “passed the exam” for the machine.
There were several attempts to pass the test, but without any particular success rate. Turing has calculated that by 2000 we will have the technology that an average person will be able to distinguish from people with a probability of 70%.
In 2014, a chatterbot named Eugene Goostman managed to convince customers that he was a real man, with a 33% success rate. The media dealt with this issue, stating that it was the first program that managed to pass the Turing test, which cannot be fully verified. The software allegedly spoke to a teenager from Ukraine, so they were able to hide the mistakes more easily.
Today, however, we are in 2018, when artificial intelligence managed to convince two individuals to belong to Homo Sapiens. So? Is it really the first software that passed the Turing test?
Officially not yet! According to Wikipedia, the test should look like this:
The assessor would be aware that he was one of the two partners in the machine dialogue, and all the judges would be separated from each other. The conversation would be limited to a text channel only. If the assessor cannot reliably differentiate the machine from a person within five minutes, it is considered that the machine has passed the test.
Clearly, Turing is only a thought about computer-related conversations, but when we consider all the results, we cannot yet confirm that Google Duplex has passed the Turing Test.
Nevertheless, neither the facts nor the Duplex’s achievements can be challenged. We believe that this software will be able to manage the conversation for longer than five minutes even if it does not manage to answer certain questions. Although the test is a bit nonsense because Duplex can almost certainly handle all the criteria. This however, instantly raises several ethical and legal issues.
Robo out
Let’s start with one very simple question. Do people have the right to know they are talking to a robot? What did the hairdresser think when she found out who was on the other side of the headset? How much would we feel deceived?
There is one famous and unusual theory that claims that robots that do not look like living beings can be cool and would not cause fear in humans. In the opposite case, they become extremely frightening, like zombies. We believe that Duplex’s voice is so professional and human-like that it just belongs to the top category of robots.
After the Duplex demonstration, the comments overwhelmed the Internet. Some were horrified, many showed elemental ignorance, and then lawyers and ethologists who were thinking about the best possible solution appeared. Google ended the discussion by pledging that, if the product was ready, some kind of “introduction” would be included, which would clearly and unequivocally inform the other party what type of communication it could expect.
But this solves only a minor part of the problem. We have to face the fact that technology has reached a point where robots can speak as human beings, and no one can find a clear difference. Google has promised that a notification will be included in the system, but what if others do not?
This is a problem we have not been thinking much about, but after Google’s latest project, we might need to think about it. What if an algorithm is there like this starts to starve a doctor, prime minister, your colleague, or your mother? But the reality is the system is quite ready and just waiting for the moment to become a part of our everyday life. It will be amazing to see if it will be accepted or rejected.