State of artificial intelligence in 2019
Repeat the word many times and it will lose its original meaning. For many of us, the phrase “artificial intelligence” has long meant a much broader concept – AI is now present in almost all technologies, and controls everything from the TV to the toothbrush. The term itself has known as Internet of Things.
However, this cannot be said about the technology itself, which is constantly evolving – is it good or bad. Artificial intelligence is used in healthcare, helps people write music and books, checks our profiles, makes a decision on creditworthiness and adjusts photos from phones. In other words, this technology makes decisions that affect our lives, whether we want it or not.
It is very difficult to ignore the hype that technology companies and advertising agencies create around AI. Take, for example, the Oral-B Genius X Toothbrush – one of the many devices presented this year at the CES conference. Brush makers loudly claim its AI-based capabilities. However, having studied the press release more closely, we will understand that this is just information about whether we brush our teeth a sufficient amount of time and in the right direction. Yes, smart sensors are built into the brush that tell you where the toothbrush is in the mouth, but calling it artificial intelligence is just plain stupid.
Another problem is a misunderstanding of technology. Laudatory articles in the press can greatly embellish a product, calling any solution that has at least some distant relation to AI an innovation technology. Very often, this is accompanied by a lack of understanding of what exactly is artificial intelligence. For ignorant people, this can be quite a serious problem, and they can combine modern AI with a more familiar version – the idea of a computer many times smarter than a person can. Specialists call this special kind of AI general artificial intelligence, however, if something like this is ever created, it will not happen very soon. And until then, it is unlikely that someone will be helped by exaggerating the capabilities or intelligence of AI systems.
Thus, now it is preferable to talk about “machine learning” rather than about artificial intelligence. The technology of “machine learning” is a section of AI, covering many methods that have a great impact on the world today (including the so-called deep learning). This phrase does not carry the mystics of AI, but more accurately explains the capabilities of technology.
The main advantage of the method is obvious – there is no need for programming. Of course, you will have to perform a large number of updates, optimize the information processing system and find more convenient ways to absorb this information, but the system does not need to say what it is to look for. That is, a machine can notice patterns that a person may not pay attention to. And since the program needs only data – units and zeros – it can be trained in a huge number of tasks, because the modern world is full of these data. If you imagine that machine learning is a hammer, then the digital world is full of nails ready to be driven into place.
Nevertheless, do not forget about the shortcomings. If we do not train the computer, how do you know how it makes decisions? Machine learning systems cannot explain their way of thinking, therefore, the algorithm may not be used for the best purposes. In addition, since the computer knows only the data that has been downloaded to it, it may form a misconception about the world or it may be able to perform only narrow tasks similar to the data you’ve previously viewed. The machine does not have the common sense inherent in man. You can create the best cat recognition program in the world, but it will never tell you that cats should not ride motorbikes and will not determine the name suitable for the cat.
Teaching computers to learn on their own is a brilliant solution. And, like any decision, it involves some compromises. Yes, there is intelligence in AI systems. However, it is not innate and plays by its own rules. This is the same as asking: “How smart is this book?” Or: “What experience does the pan have?”
So at what stage is the development of artificial intelligence? For several years (and so far), the headlines have promised us another technological breakthrough, but now some experts believe that we have reached a certain milestone. However, this does not mean that the development of technology has stopped. From a scientific point of view, there are a large number of areas for studying inside existing knowledge, and from a product point of view, so far we only see the tip of the algorithmic iceberg.
While artificial intelligence – machine learning – is still something new, often inexplicable and not sufficiently studied, however, in the future this technology will enter our lives so much that we stop noticing it.