Artificial intelligence causes nightmares for people

A group of scientists made a new idea on artificial intelligence, they loaded 200,000 images of human faces into a computer and then inserted a zombie image to create human nightmares.

Today's artificial intelligence mainly learns good things from people, they mimic human thoughts and accomplish tasks by repeating, or a little more. But this does not make people less afraid of the fear of artificial intelligence being self-aware and dominating the world.

Many people believe that artificial intelligence is not really scary, need to change to make it more frightening. So researchers from the Massachusetts Institute of Technology and CSIRO have built artificial intelligence that makes it possible to create human nightmares.

By helping artificial intelligence understand horror images, it will act in reverse to help people feel safer.(Photo: DeepDream Project / Google).

This computer is called the Nightmare Machine , its algorithm is also to create nightmares. Researchers loaded 200,000 images of normal human faces to recognize and create images of normal human faces.

After that, they loaded into the zombie image so that the machine could understand that it was a scary image and would create such scary faces from the normal face image. From here the machine will create terrible nightmares.

However, creating nightmares is not the main purpose of this machine. It has a higher purpose, creating nightmares to understand fear and comfort people. When the machine learns things that are frightening, it acts in the opposite way to create behaviors that people feel safe and comforted.

"Just like children, by learning to harass people, this machine can be trained to avoid those behaviors. Therefore, this technology may seem silly, but it will really be useful in making people feel comfortable, "said Manuel Cebrián Ramos from CSIRO.

When the machine learns things that are frightening, it acts in the opposite way to create behaviors that people feel safe and comforted.

Mr. Ramos believes that this is the best way to create peaceful artificial intelligence machines because it will behave against the horrible things. Instead of setting rigid one-way rules, we can apply the opposite to achieve greater efficiency.

Basically, we first make machines understand us, then we will work with machines easily."This is like for aliens, if we understand each other, we will work well together. Otherwise it will easily lead to non-cooperation and conflicts and conflicts," he added. .