14
Jan
2019

Developer Explains: What is Artificial Intelligence?

Every day we can read news about modern phones, speakers, headphones, or washing machines that allegedly contain artificial intelligence (AI). Do we really have thinking machines already? And if not, what does AI actually mean? Raimo Seero, Chief Technology Officer of Uptime, published a thorough overview in the daily newspaper Eesti Päevaleht (in Estonian) in December, explaining how developers see the software that is referred to as artificial intelligence.

For developers, software solutions which are called artificial intelligence are not actually thinking machines but use machine learning. Since computational power has increased very much thanks to new technology, a large amount of data can be fed to a computer, which can then look to find connections and logic – a kind of massive processing of data that is far beyond human capacity.

Uptime has experience with several AI developments in the field of taxi services and media. In some situations, learning algorithms are already pretty good and they can be applied to get a better result if other tools don’t work. In such cases, when it would be reasonable to implement machine learning, we can find a suitable solution for clients.

There is still a long way to go (but AI is already very effective on some occasions). Artificial intelligence is also said to be present in digital assistants and chat bots used by larger companies in customer service today. Are the machines really thinking or is it only a big marketing bubble?

Answering that question, we must admit that AI is not a specific technology – 10 years ago there were also text-to-speech robots, and now robots can recognize objects from photos, for example. The term AI covers these, and many other different solutions and its meaning are constantly changing.

From the software developer’s point of view, all that is now called artificial intelligence is, in fact, only a part of the software solutions that reached new levels in recent years, which can offer new exciting solutions thanks to the rapid development of cloud technologies and algorithms. For example, artificial intelligence is associated with data science and processing of big data, machine training, neural network based algorithms and so-called blurry logic (logic as the human brain thinks, not black-and-white true or false).

The software can provide answers so naturally that it may seem like the machine thinks for itself. But behind all these smart solutions are programming methods, which can be much more complicated than before, but are not yet directly related to a self-thinking system. Artificial intelligence, capable of learning, developing and making independent decisions in any field, is still far from existing. We are not sure how far away that really is – years, decades or maybe even centuries.