Everyone who watched the Terminator franchise knows a little bit of artificial intelligence. We are nowhere near that level of technology, though computers are getting smarter by the day. Artificial intelligence means a machine capable of reaching its own conclusions and having a thought process, much like a human. They should be able to perform tasks which humans do, to the point of barely being able to tell the difference.
Artificial intelligence is one of the leading technologies which is changing our world, and here is how.
Scripts vs A Real AI
When people think of an AI, they either go full Skynet, which is impossible at the moment, or they think of an AI from a video game, let’s say a real-time strategy. These AI are not actually AI, but scripts. They perform some tasks which they were previously assigned, depending on a couple of conditions. Often, increasing the difficulty in these games would only result with the opponent getting their stats buffed, or their speed of construction and unit training increased. This is not really a better AI, just one which can do things faster, through what is basically a cheat.
Real artificial intelligence would have to be able to make its own decisions based on the input it currently has, not to resort to scripts. Better executed computer opponents seem much more human-like, because they are following more complex instructions. A great example of this was the testing of a new AI in League of Legends, which was done on the Brazilian server. There are clips of an AI playing better than high-elo players on a couple of occasions.
Problems with AI
Artificial intelligence runs into its own set of problems. For example, for an AI to become better in a certain task, it requires huge amounts of data. This is not a novel topic, but for some use cases, the data requires is just not there, or is difficult to feed properly, as the quantity of data might not have the necessary quality.
AI has a couple of other problems, which were narrowed down to knowledge representation, and in this category, specific sub-categories named default reasoning, sub-symbolic thinking, also known as intuition, and common sense. Other problems include planning, learning, motion, social skills and social intelligence, intelligence in general, perception and even language processing. AI has a long way to go before it reaches Skynet status.
Applications of AI
Some of the best applications of AI are in medicine and healthcare. Microsoft is developing an AI called Hanover, which is supposed to help doctors decide which treatment is best for cancer. There are so many treatments and medicines, that doctors often cannot conclude which one is the best. An AI could go through all the papers, research and combinations much quicker, presenting optimal results, which the doctors can then review.
AI in video games is familiar, already, from soldiers who can find different paths to subdue a target to normal people who go to work everyday, but also react to you in an appropriate manner.
In the automotive industry, artificial intelligence already has some basic functions like changing lanes and slowing down or speeding up, or driving on a highway. This industry faces many safety concerns so AI development is very important and is taken seriously.
AI is changing our world, slowly but surely. As long as it doesn’t create killer robots, we should be fine.