Can a Hollywood movie in this series help us deal with the big problems with its technology? artificial intelligence; The Greek public and critics agreed that the film “M3GAN” it’s a pretty boring thriller. However, putting aside the gimmicks of horror films and the overacted script, we are faced with some of the most important challenges of achieving superintelligence in decades that we must not forget.

The film may not deal with them in detail or resist the lure of sensationalism, but director Gerard Johnston’s intentions to put some of the most classic problems of artificial intelligence into a pop culture version of it are remarkable.

The young heroine of the film loses her parents in a car accident. Until we get self-driving cars that eliminate human error, humanity will continue to mourn the victims on the pavement. The very manner in which the conflict occurs speaks for itself. The car is in such a state that no matter what the driver does, the result will be the same. OUR technologies it’s not neutral like the movie says. They are trapped in circumstances. This must be remembered before embarking on the supermind. A little girl is cared for by her mother’s sister, a successful employee at a leading toy company. When the new guardian realizes that the little girl has a hard time adjusting to new circumstances, she gives her her own invention, the M3GAN robotic doll, who becomes her best friend. Her employer is excited and wants to put the doll on the market as he is sure it will be a hit. The film deliberately exposes the emotional immaturity of the new mother.

With the robot, she solves not only the problem of her job, but also the problem of fatherhood. The robot doll is a logical extension of the iPad nanny. Gradually, however, we discover that the solution was temporary. The robot doll has a life of its own. The updated legend of Frankenstein is back, this time in the nursery. Sophisticated AI robots gain some sort of consciousness, will, and intelligence. As a result, M3GAN refuses to comply.

When the human factor has to intervene, it acts with its head, based on the initial instructions received. Thus the nightmare of HAL in Kubrick’s 2001: A Space Odyssey is resurrected once again, and Stephen Hawking’s anxiety is realized. The task of a smart machine is to satisfy the little girl as much as possible and protect her. When a little girl is threatened, M3GAN goes so far as to commit crimes. The film, among other things, is dedicated to the so-called “paperclip theory”. If we gave the superintelligence the goal of maximizing the production of paperclips, the machine would take its goal literally and exhaust all the resources that fill the planet with paperclips.

I would like to hope that when the question arises of bringing such machines to the market, our reflexes will be strong.

So being able to turn it off whenever we as humans see fit is fundamental. Beyond the need to be able to push off, the movie reminds us of what AI experts have called the “leveling problem.” If we can’t teach machines our own moral codes, our own complicated moral imperatives, we’ll run into similar stories about Bartleby robots. However, even if we find a way to deal with machines refusing to comply, attackers do not exclude the possibility of exploiting security holes or creating their own uncontrolled and unethical machines.

However, the ending of M3GAN is perhaps the film’s most unsettling element. While we are freeing ourselves from the course of action and after the catharsis has occurred, the last frame focuses on the device of the virtual assistant living in the family kitchen. This could be an Amazon, Apple, or Google device. This device is here today and already knows a lot. Soon he will learn much more. She followed the plot of the story, she knows everything, even this text will someday become her property, so the director dedicates her last frame to her, telling us that artificial intelligence is already here and it is taught daily a huge amount of information that we voluntarily provide to her.

Virtual assistants are gathering vast amounts of knowledge and data, making machine learning models ever more powerful. The question is, have we been able to fine-tune this powerful technology.

The M3GAN catastrophic scenario is not extreme. Anyone who has seen Boston Dynamics robots understands that the possibility of building machines like the robotic doll is obvious. The software is still to be improved. I would like to hope that when the question of bringing such machines to the market arises, our reflexes will be strong, and we will already take the necessary active steps. Otherwise, we risk turning a cheap horror movie into the center of our storytelling.