Whereas in 1990s the Terminator’s storyline seemed as far from the reality as A Nightmare on Elm Street, now the technological anti-utopia has become a serious topic for discussion. AI developers have made a great progress in the hominization of the computer mind. For instance, in 2015, three scientists from Russia and Ukraine presented a chatbot named Eugene Goostman that beat the Turing test for the first time ever.
The Turing test defines whether a machine can think. During the test, a person communicates with a computer and a human. The person does not know who is who and has to define it basing on answers. Eugene Goostman tricked more than a half of its companions making them think that they were talking to a human.
If three developers have managed to cheat the crowd, what will happen if governments arm with AI technologies making them a part of everyday life?
Positive forecasts say that human life will become significantly simpler, as robots will release the society from the routine. However, there are also gloomy hypotheses. Read further in the article which harm AI can do and how to resolve this issue.
The modern industrial economy already partially depends on the artificial intelligence, but the long-awaited automation of routine work can increase the number of unemployed people.
For instance, the stable financial situation in the USA depends on consumer loans. They are serviced by AI solutions that check applications for credit card issuance, issue permits to make payments, and detect fraudulent transactions. As a result, credits are arranged quicker and at a lower price, as the procedure that was once carried out by a whole team of employees now requires just a couple of people and one program.
It leads to the reduction of workplaces. English scientist and futurologist Arthur Charles Clarke, known for the work on the 2001: A Space Odyssey science fiction film produced by Stanley Kubrick, wrote, “People will face the future full of terrific boredom. The main problem of their lives will be the choice from several hundreds of TV channels”.
But boredom is not the only problem of the future. The 2017 report of McKinsey&Co consulting company shows that robots will make 800 million people unemployed by 2030.
Increased government’s control
Such works as Equilibrium, Nineteen Eighty-Four, the Maze Runner, and Black Mirror show that people fear the strict control from the government. Now this fear looks quite justified.
For instance, China plans to introduce a social credit system by 2020. The system with embedded AI will define the credit score of citizens by monitoring their social behavior: how they spend money, how regularly they pay off the bills, how friendly they are with other people. The trustworthiness of every Chinese citizen will rely on this assessment. Those people that do not reach the required score level will not be able to leave the country or use some public services.
However, the interference of artificial intelligence in our everyday life may serve for good as well. For instance, a camera with the face recognition technology detected a thief that had been missing for several years in the Russian subway.
Although AI technologies offer a high speed and quality of data processing, they are not always reliable. After all, people develop machine learning and they sometimes neglect moral and ethical principles in the process of product development.
Even programs produced by Google, the leader in the use and development of artificial intelligence, make errors. For instance, Google Photos service mistakenly detected Asian people as blinking on the photos.
In the USA, artificial intelligence is used in the law-enforcement and judiciary system, which sometimes causes scandals. For instance, Northpointe software, which analyzes crime statistics and gives recommendations, showed a strange tendency: it believed that dark-skinned people were more prone to commit crimes than white people. The bug was fixed. However, representatives of Northpointe did not comment why the race was included in the list of criteria that influence the amount of crime.
“Start with yourself”: teaching tolerance to AI developers
Artificial intelligence is not the author of racism and does not dismiss people from work. Although many computers can take decisions, it is just an imitation preset by programmers. AI just chooses an appropriate algorithm based on definite specified parameters.
Big Brother is a close and (in case of China) quite possible prospect, but still it can remain a science fiction story. Modern scientists, developers, and entrepreneurs work to make artificial intelligence secure for the society.
Rapid progress of the artificial intelligence gave birth to a number of initiatives focused on ethics in the AI training. Some of them are OpenAI organized by Elon Musk and Sam Altman, Partnership on AI, research center studying the ethics of artificial intelligence at Carnegie Mellon University, and ethics department at DeepMind. They regularly organize summits, forums, and meetups among specialists related to AI. Their goal is to direct the technology development for the good, find problematic places, and form general rules of artificial intelligence creation.
How AI will reach zen
The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems has been working on Ethically Aligned Design since 2016. It is a set of rules and recommendations on the work with artificial intelligence. To add the non-Western value systems to the document, the initiative established a Committee for Classical Ethics that studies the influence of Buddhism and Confucianism philosophy on the work of artificial intelligence. The final version of the document will be ready by the end of 2019.
As we can see, the behavior of artificial intelligence rests in hands of developers and entrepreneurs that create it.
Isaac Asimov, the writer of science fiction and author of famous Laws of robotics, wrote, “If artificial intelligence is programmed to be indifferent to people, it will lead to a disaster. We should make sure that AI is secure if we want to survive its emergence”.
Learn about the subtleties of development and use of ethical AI at AI Conference that will take place in Moscow on November 22.