Due to the recent success of neural networks in areas (drawings, music, programming, etc.) where they were considered incapable of succeeding, the letter AI is now attracting everyone's attention 
By the way, let me try to guess - your example is about processing statistical information? It seems to me the most obvious, because knowing the "digitized personality" of an athlete/team, you can more accurately predict its results.
Was plenty surprised when a lot of AI samples generated realistic images. Not realistic drawings, but realistic images. I wouldn't have doubted it was a photo taken by someone else really if I didn't see that it was just AI-generated, I feel like just a few months back I doubted how AI can do this and that but now they've pretty much done a lot already comparatively speaking. By the way, let me try to guess - your example is about processing statistical information? It seems to me the most obvious, because knowing the "digitized personality" of an athlete/team, you can more accurately predict its results.
Also, I'd be plenty surprised if there was some sort of method that didn't need statistics to generate a guess. Generally speaking, most predictions made by people often rely on data, it can come from a lot of sources though and some may be incomplete or not but that's basically the basis for everything. Tbf I reckon an AI built on top of how bookmakers provide their odds can probably give out semi-accurate predictions already.
This was inevitable as the AI makes its generations of (whatever) better and better. For example, it is quite difficult to distinguish a fully generated person from a photograph of a real person.
As for predictions based on data, the real breakthrough will happen when the AI itself begins to collect them, and not just use the data sets provided by people/game protocols, etc.
