AI – real or a dream ?

In the 80s when Japan was at its peak and software languages were third generation, the Japanese announced that they were skipping the fourth generation (I’m not sure what it was supposed to be) and going straight to the fifth generation which includes artificial intelligence. The whole world was shaken, the Americans fought back and invested huge sums, there was a feeling that artificial intelligence was coming. Later disillusionment set in and the field faded and became relegated to academia and industry. The development of computers (computing power) and the development of new algorithms created the next generation and today there is renewed hype in the belief that this technology is going to change the world. Is that so?

On the one hand we see the development of the field, not only on paper, but real applications with amazing products and companies that invest huge resources in the field. On the other hand, the products are still immature, with tough problems of product control, reliability and quality, and there is a difficult question as to whether it will be possible to solve them in the future. Let’s take for example the quality of the products. Those who have played with ChatGPT know that sometimes the answers are really stupid or wrong. The main reason stems from the fact that the system is based on existing information written by a person that is not always correct (for example ‘the Earth is flat’), so there is a non-negligible chance that an answer will contain incorrect parts. The issue gets more complicated since the information in the future will be written in part by the technology itself which, as mentioned, contains incorrect components and therefore the element of unreliability in the answers will only increase. Worse, there is a chance that the content writers will prefer to use content generated by the system and not invest in original content, thus increasing the unreliability of the answer. And I haven’t mentioned product control, and a number of other complex issues.

When a plane advances to its destination, there is a ‘point of no return’, from that point the plane does not have enough fuel to return and must continue to the destination, even if there is a need to return. Every technology has a point of no return from which it will continue to develop without the possibility of going back – we cannot stop using computers, go back to riding donkeys or living in caves. Artificial intelligence has not yet passed the point of no return, there are still not enough uses or enough users, the hundreds of millions of users currently mainly playing/learning the system. To accumulate more uses and users we must solve (at least partially) the essential problems that exist in the technology, and this takes time. Today the field attracts investments and entrepreneurs, but there is a chance that the hype that exists today will pass before we pass the point of no return and with it the investments will disappear. If this happens, inherently conservative systems (such as the education system) will be the first to stop the activity in the field and prefer to not even try.

Now to the main question – will artificial intelligence cross the point of no return. In my opinion, the signs are promising – the leap forward made by technology, the applications that already exist, the investments made in the field, the public’s interest, and the existing potential – all indicate that there is a good chance that within about a decade artificial intelligence will reach a level where it will be impossible to live without it. When it reaches this point, the education system will also be forced to get on the wagon (see AI – an inevitable technology).