If anyone was a teenager like me in the 1980’s, they probably remember the movie „wargames“ and the ending of the film in which the computer learns through repeated games of tic-tac-toe that some games are not to be started as they cannot be won.
That was fiction. At that time, and up to now, the most powerful computers we built, with ever-increasing power and speed were prisoners of the functions we assigned them. However, I read recently that „DeepMind“, the AI entity tested by Google is programmed and continuously improved to learn: not only does it not forget previously learned skills, it also finds a way to use them in new activities it undertakes.
And only recently, „DeepMind“ has been in the news for playing video games, and given the circumstances, it chose either to cooperate or to be aggressive in order to win.
So human: choosing aggressiveness in times of stress, or ready to share when it reserves the best outcome. A very interesting turn of events, because I am not sure when we are discussing about the breakthroughs of artificial intelligence that we really foresee the capacity of the entity to develop traits close to those of human beings.
When we read about the new technologies enriched by AI, I find that we mostly focus on the benefits of a more powerful tool that will make better decisions based on its ability to process much more data.
We hear about the activities that could be replaced by AI: salespeople, front-office bank clerks, customer assistance. It is even happening now on stock markets as more and more decisions are algorithm-based and less based on the intuitive human approach of traders. That is what a tool does, but with the latest developments, are we still talking about a tool?
We rarely think of a technology that could one day match the human mind and interact socially with others. We can easily imagine AI organizing tasks, providing leads to salespeople, measuring performance, and identifying high-potential assets. In a more complex approach, if the entity would go beyond these tasks, could it learn enough about the human mind and find how to communicate with members of the team, how to motivate them, to encourage them to grow? The jobs replaced might not be the only ones we think about today.
Of course, we are not there yet by far. More generally, it is in my opinion a vivid question that need to be taken in account in the communities in which we live. If we develop artificial intelligence with a learning curve, it will grow by definition, it will learn many skills but also it will also learn to know us. What we are going to do with this technology will be linked to how we choose to live as a society.