Whether you are a college football fan, an NFL enthusiast, a fantasy football player, or a mix of all three, you will likely read a lot of content based on analysis and predictions for the new campaigns ahead. You will, of course, be familiar with the type of content we mean: Win-loss predictions articles for the season, player projections, new season power rankings, and so on.
Most certainly, some of these articles are interesting. The best ones will use painstaking research and data to at least make a solid argument about how X or Y could happen across the college football or NFL season. The majority of us will look at those prediction pieces objectively, allowing them to complement rather than shape our own opinions. Just because an ESPN analyst says Ohio State will win the championship doesn’t mean we have to agree.
Season projections are a tricky business
And yet, it is interesting to note that there is a new phenomenon around this time of the year – AI prediction articles. This type of content has been around for many years, with titles like “MIT Scientists Use Supercomputer to Predict March Madness,” but they have become much more prevalent since the arrival of ChatGPT and similar tools. Often, they are harmless, and those who are experienced in using stats for sports betting and fantasy games will take a skeptical approach, regardless.
However, there is also something of a leaning toward AI as some kind of soothsayer or omniscient presence. You can see this outside of sports. For example, on X, the role of Grok (Elon Musk’s AI chatbot) has become something of an arbiter of truth, with users asking the bot to verify information and even settle arguments. It all builds into a conception that AI is both smarter and less biased than the average person. This is not true – yet.
But in terms of sports, it’s important to understand what AI is doing when it is asked to predict the college football season or NFL playoff brackets, or whatever else you want it to churn out. The bot doesn’t suddenly watch 1000s of hours of Texas Longhorns games, analyze every play, and give you an answer on the team’s season: It will spin around the internet, look at elements like college football projections, NFL Playoffs odds, and give you a summary of what that content says dressed up as a specific answer.
AI can only mimic human opinions
Almost immediately, you can see the problem. The AI predictions are presented as thoughtful answers, when in reality, they are regurgitations of information already available on the internet. AI doesn’t have an ‘opinion’ on Notre Dame this season, but it can look at some analysts who do and give you an answer that looks like an opinion. The other problem, of course, is that there is a limit to what AI can access. It certainly cannot easily access the expert data that specialists will use to make sports predictions and set betting odds.
As we said earlier, a lot of these articles are fairly harmless. They are not meant to be taken as expert advice for betting or fantasy sports. Indeed, you should note that a lot of these articles pop up on non-sports-specialist websites and publications, so they are not targeted at professionals. However, if you look closely underneath the hood, you will see that they are fairly two-dimensional. AI might one day be the tool that dominates sports predictions, but it is not yet ready.
