I've been participating in a beta for MidJourney, an AI image creation algorithm that takes key words, searches the internet for thousands, millions of images and creates an original graphic. The results are dream like, strange, and sometimes really well composed.
There are many AI image making systems in development, for example Artbreeder and Dream are ones I've dabbled with. This technology is just in its infancy and is already exciting, intimidating, and a little scary for those of use who make a living doing Illustrations.
As an experiment I challenged the software to replicate the style of David C. Sutherland III. The results are interesting, but not particularly convincing. The images are more abstract than illustrative, and they don't really capture the fine line work at all. I've found in general it is difficult to get the AI to draw decent line work, it tends to do either painterly or photographic much better. Here are some of the results.
This was only a first attempt. There are a lot of techniques such as refining keywords, emphasizing words, feed it images to work from, and so on that can help to improve the results. The keywords I used for this batch were: knight fighting a dragon in a dungeon, and a paladin in hell fighting devils. As you can tell the answer to the question "Can AI draw like David Sutherland?" is currently no.
I'll be continuing to play with the software as it improves, but for now illustrators don't have much to worry about. Then again, master Go players never thought they'd get beaten by a computer either. We live in interesting times.