 What does this painting, this book, and the background music you're hearing right now all have in common? They were created by robots. For years now robots have flown planes, written news articles, and even controlled stock markets. As automation has replaced menial and repetitive tasks, the creatives of the world have laughed at how easy it is for robots to replace the common working man. But it would be impossible for a robot to replicate their style, abstract concepts, and deeper meaning. Well, maybe not. So who's laughing now? Modern computers were first invented in the late 1960s and since then their computational power has risen exponentially. The computer that used to fill an entire room is now less powerful than microchips the size of a dime. With this explosion in computational power, computers have been able to move into ever more technical tasks. From the early days of simply moving between predetermined paths on automobile floors, robots can now maneuver complicated spaces and even develop medical breakthroughs. But these advances are based on very tangible criteria whose effects can be easily quantified. Moving around is really just going from point A to point B based on what's around you and computational medical breakthroughs come down to loads of numbers and statistics, all things that computers are good at. But these tasks have concrete goals and very easy ways for a machine to determine success and failure and to see how it's performing. A good example is the YouTube algorithm. Its simple goal is to increase watch time across the platform. So it shows different users different videos and learns from what works and what doesn't. If people watch a lot of a video, the algorithm decides to show that video to more people. If a lot of people stop watching a video, the algorithm will let it fade away. Well, it might be a lot of data. The general process is fairly straightforward. But how do you teach a computer the intangibles? When a person looks at a beautiful painting, they may feel a rush of emotions. But to a computer, it just sees a handful of pixels. And if a computer struggles to appreciate art, how can it recreate it? Despite all of the art coming out by robots, most of it isn't really art. At least not what we would consider art. Most paintings made by AI come in three varieties. The first, very simply, a robot will be given an image like a painting by Picasso or look at a live model and paint what it sees. The most common technique involves the computer comparing its canvas to what it wants to recreate. And then it changes the part of its canvas that looks the most dissimilar. This way, over time, the painting looks more and more like the subject until it's done. This isn't really art in that there's no creativity behind it. These robots are basically glorified cameras. Another common technique programmers have used to create artistic robots is by having them copy styles from existing artists. These programmers will feed a computer an image like a Picasso, and the computer will analyze the style of the artwork. Then, the programmers will give the computer a new image, and the computer will alter the image to appear in the same style as the first image. Again, while this can create a unique artwork, the programmers are just teaching the computers to rip off existing art, not to create their own. A more recent style of robotic art is one where the computer is given an abstract set of algorithms to follow. Think of these algorithms as a very rough set of rules to follow, whereby the robot can make original decisions. Variations of this last style would have brought robotic artists into the spotlight in recent years. As the algorithms the robots follow have become more complex, and neural networks and deep learning have allowed robot artists to teach themselves new designs, these robotic artists have begun to make their own styles and leave their footprints on the art world. While a recent social experiment found that art experts could differentiate between man-made and robot-made art, the vast majority of the time, that doesn't mean that one is necessarily more valuable than the other. Back in 2018, a portrait titled Edmund de Bellamy, which was created by an AI, sold for more than $400,000 at auction, and robots influence over their artistic world isn't just limited to paintings. For years, many news organizations have used AI to create new stories, and these same AI are being adapted to create stories. And Spotify is suspected of working on an AI to generate free generic music to save on royalties, after they're telling hire François Pachette from Sony, who had been working on AI software to develop music. As artificial intelligences continue to advance, and they continue to incorporate deep learning into their algorithms, robot artists will only improve and may even force us to question how we consider art. The rise of machines may even leave human artists in the dust. Be sure to subscribe and click that notification bell so you never miss an upload. And remember, there is always more to learn.