Computer

The first artificial intelligence in history was a fiasco

Not many people know it, but the first artificial intelligence it was invented in the 1960s by a psychologist named Frank Rosenblatt and it was a colossal fiasco.

Rosenblatt has dedicated himself for years to study of cognitive systems and had set himself the goal of replicate the functioning of the brain human. His most significant project is called Perceptron and today it is considered the first example of intelligence artificial of history. Even if, at the time, he disappointed the expectations of the public and professionals.

Chi era Frank Rosenblatt

The profile of Frank Rosenblatt it’s not easy to frame. Born in New Rochelle in 1928graduated from Cornell Universityand then start working on Cornell Aeronautical Laboratory.

Psychologist and researcherRosenblatt would gradually specialize in the field of cognitive systems. And it was at the Cornell Aeronautical Laboratory that he began working on Perceptron: a computer that is today considered the first ancestor of ChatGPT.

Il first Perceptron hardware was completed between the 50s and 60s of the twentieth century and can be considered a real precursor of artificial intelligence.

The system was in fact capable of learn e di develop new skillsthanks to one neural network conceptually not too different from the current ones.

Perceptron learned simulating the mental process of the human being, through a complex trial and error mechanism (trial and error).

The failure and success of Perceptron

The expectations on the Mark I Perceptron were very high: some journalists presented him as a real one rival of the human brain.

From a certain point of view Rosenblatt had anticipated not only the perimeters of intelligence artificial as it is understood in this historical moment, but even those of thegenerative artificial intelligence.

Perceptron should have been able to not only carry out tasks independentlyor to translate input from one language to another in real time but, according to Rosenblatt, it even could have develop original ideas.

I results but they were rather resigned. In fact, on the first test Perceptron was only able to distinguish the position of the brands that characterized certain cards and he succeeded after about 50 attempts.

Today, however, the work of Frank Rosenblatt is considered of vital importance for the development of artificial intelligence contemporary. The same Cornell University has dedicated part of its official website to memory of Perceptron.

In fact the single layer neural network of the system represents the basic structure of AI tools based on machine learning and deep learning. For this reason it development of Perceptron is often compared to laying of foundations on which artificial intelligence and generative artificial intelligence would then be born.

To know more: Artificial Intelligence: what it is and what it can do for us

Leave a Reply

Your email address will not be published. Required fields are marked *