What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Will AI Replace Human-Sourced Fantasy Football Content? (1 Viewer)

AI will do everything in the next few years. So yes.

In my opinion the sharks will be shifting through the deluge of data that we'll be able to throw at AI to find what data points are relevant and what aren't, while everyone else is caught up in what is most reshared.
 
It will start being used for projections and lineup calls very soon. We are also likely not far off from an AI operated team drafting and possibly managing a team all season, which is not to be confused with auto draft.

My guess is if AI is successful at these things we'll either have a great new source for info or need to find a new hobby.
 
Is AI creatively thinking or just regurgitating what's available? I've messed around with ChatGPT and what's spit back looks like a conglomeration of the web. We can't agree on the impact of weather on games. Is there enough specific data available and sample size to solve this riddle and then create an accurate weather forecast for a game. It would probably take 10 years to prove out AI is better at FF and that would require the AI to remain static.
 
I think it will. I’ve been gathering and reviewing past drafts from league mates for over a decade and have been decently successful at predicting picks during a live draft. If AI got involved it would be way better at it. Also, check out this presentation from the guys who did social dilemma. You will be both amazed and horrified.

 
AI talk is over blown. I will believe it when I see it. And I'm a software engineer

Boy I dunno. Some of the things it can do now are amazing, and it's getting amazinger with every iteration. At some point in the not too distant future I can see it picking up nearly all data analysis in just about every industry.
 
Well I should have read Faust's article first. The AI basically plagurizes the Innernets. Sure this is amazing tech if you're a high school or college student and need a history paper written, it's not so good if you're trying to determine if you should start Russel Wilson this week and the AI has honed in on Denver homer as a source of information.
 
AI talk is over blown. I will believe it when I see it. And I'm a software engineer

Boy I dunno. Some of the things it can do now are amazing, and it's getting amazinger with every iteration. At some point in the not too distant future I can see it picking up nearly all data analysis in just about every industry.
So where are you getting good data analysis in FF? In theory for AI to worthwhile, it would have to check the accuracy of the data and I'm not sure the sample sizes are large enough to distinguish between accuracy and luck.
 
AI talk is over blown. I will believe it when I see it. And I'm a software engineer
Thanks for sharing your opinion.

I just can't wrap my head around AI overcoming a developers bias. In the case of fantasy football to get a good result, AI would need to analyze not only the web, but every TV show, radio show, and social media platform and then accurately weight that information. How it's going to differentiate between the guy who started the Anquan Bolden thread in early August and the guy who started the Marshall Faulk ball cancer thread?

You can wind up a toy, but when you put it down it's still has to face some direction.
 
AI will absolutely dramatically impact fantasy sports. It's almost the perfect industry to be disrupted.

I live in the AI world so I may be a bit biased, but I don't think there's any question about this. Just look at the evolution trends of ff analysis. It is inevitable.
 
Last edited:
As someone who works in the CG VFX industry, I can say that AI definitely will, and already is, having a huge affect. Generative art, learning to assist in some of those most mundane tasks (creating roto mattes, painting out light stands or tracking markers), taking small subsets of data and giving you the ability to fill in the rest (both in terms of video “in betweens”, or 3d geometry creation from a sparse set of images,or image upressing), or plate extensions and full cg environment creation, or texturing 3D geometry, or creating variations on creatures to fill background crowds.
and notice I haven’t even mentioned deepfakes yet.

it’s a big deal for sure in the entertainment industry. Questions around “ownership” and copywrite come into question for AI generated content, mostly with concerns around what datasets were used to train these models. And one of the big sticking points in the current writers strike is AI generated screenplays allowing studios to have AI generate piles of “stuff” to help inspire new stories, With the biggest concerns being AI replacing a writer entirely. And the likely upcoming actor’s strike will probably include huge concerns around ownership of “likeness” and any datasets of themselves that might be collected for a project, with AI face replacements being all the rage.

it’s a big deal.
 
AI won't be able to take over player/prospect evaluations.

From an eye-test perspective, I agree, though perhaps in the medium/long run even that aspect is at play the way AI is evolving.

From a numbers analytics perspective, AI is going to take over very soon, and I'd bet the savviest of orgs are already using it plenty and planning for rapid expanded deployment.
 
AI talk is over blown. I will believe it when I see it. And I'm a software engineer
Thanks for sharing your opinion.

I just can't wrap my head around AI overcoming a developers bias. In the case of fantasy football to get a good result, AI would need to analyze not only the web, but every TV show, radio show, and social media platform and then accurately weight that information. How it's going to differentiate between the guy who started the Anquan Bolden thread in early August and the guy who started the Marshall Faulk ball cancer thread?

You can wind up a toy, but when you put it down it's still has to face some direction.

I think you are underestimating the machine-learning aspect of this technology. Not saying AI is going to take over in 2023, but as it becomes massively more deployed over the next 3-5+ years, it will learn how to do what the industry is now doing with an army of decentralized humans, but at the speed of light. That's if its military-AI cousins don't kill us all first.
 
AI will do everything in the next few years. So yes.

In my opinion the sharks will be shifting through the deluge of data that we'll be able to throw at AI to find what data points are relevant and what aren't, while everyone else is caught up in what is most reshared.
Won't the AI figure out which data points are relevant and which aren't?
 
AI will absolutely dramatically impact fantasy sports. It's almost the perfect industry to be disrupted.

I live in the AI world so I may be a bit biased, but I don't think there's any question about this. Just look at the evolution trends of ff analysis. It is inevitable.
The only working against it is the anti-nerd faction of football. Many are resistant to analytics. Those people are just as likely to be resistant to AI.
 
AI will do everything in the next few years. So yes.

In my opinion the sharks will be shifting through the deluge of data that we'll be able to throw at AI to find what data points are relevant and what aren't, while everyone else is caught up in what is most reshared.
Won't the AI figure out which data points are relevant and which aren't?
I'm not sure the sample size with a 17 game season is enough, especially when one in game coin flip decision can impact the rest of the game. Then you throw in new wrinkles like the QB rugby scrum or an NFL rules change and the models have to be reworked with limit samples once again. Lastly, you have coaches doing dumb things that I'm not sure AI can account for.
 
AI will do everything in the next few years. So yes.

In my opinion the sharks will be shifting through the deluge of data that we'll be able to throw at AI to find what data points are relevant and what aren't, while everyone else is caught up in what is most reshared.
Won't the AI figure out which data points are relevant and which aren't?
I'm not sure the sample size with a 17 game season is enough, especially when one in game coin flip decision can impact the rest of the game. Then you throw in new wrinkles like the QB rugby scrum or an NFL rules change and the models have to be reworked with limit samples once again. Lastly, you have coaches doing dumb things that I'm not sure AI can account for.
if AI can’t find it, how are people?
 
AI will absolutely dramatically impact fantasy sports. It's almost the perfect industry to be disrupted.

I live in the AI world so I may be a bit biased, but I don't think there's any question about this. Just look at the evolution trends of ff analysis. It is inevitable.
The only working against it is the anti-nerd faction of football. Many are resistant to analytics. Those people are just as likely to be resistant to AI.

The older crew, frankly this messageboard is the resistant crowd.

But AI will impact the analytics and content...to some degree AI has already been implemented in the sports betting industry.
 
Followup: Will AI replace sports journalism?

How could it not?
It already has to some degree. AP uses game recaps written by AI. Feed it box scores, play-by-play and historical data and can spit out a simple story as well as a human.

A lot of journalism is interpersonal skills though. Finding human interest angles, building relationships with sources, asking the right questions to get an athlete to open up, etc. AI isn't going to be very good at breaking news or feature stories.
 
AI talk is over blown. I will believe it when I see it. And I'm a software engineer
Thanks for sharing your opinion.

I just can't wrap my head around AI overcoming a developers bias. In the case of fantasy football to get a good result, AI would need to analyze not only the web, but every TV show, radio show, and social media platform and then accurately weight that information. How it's going to differentiate between the guy who started the Anquan Bolden thread in early August and the guy who started the Marshall Faulk ball cancer thread?

You can wind up a toy, but when you put it down it's still has to face some direction.

I think you are underestimating the machine-learning aspect of this technology. Not saying AI is going to take over in 2023, but as it becomes massively more deployed over the next 3-5+ years, it will learn how to do what the industry is now doing with an army of decentralized humans, but at the speed of light. That's if its military-AI cousins don't kill us all first.
I think you are overestimating these machines. They only know what they are told. Maybe dogs will rule the world
 
I think you are overestimating these machines. They only know what they are told. Maybe dogs will rule the world

And I think you are underestimating it.

AI uses code/algorithms to analyze data.

It is becoming increasingly powerful and will have a profound impact on how we live over the next century and beyond.
 
I think you are overestimating these machines. They only know what they are told. Maybe dogs will rule the world

And I think you are underestimating it.

AI uses code/algorithms to analyze data.

It is becoming increasingly powerful and will have a profound impact on how we live over the next century and beyond.
I think you are both right, but both looking at different aspects of AI. AI is a large domain and can mean different things to different people. I'm working with AI in a couple different ways currently, data processing/analysis and next generation image processing.

The data analysis piece is pretty impressive at times with algorithms, data weighing and predictive outcomes.

The next generation image processing is kind of garbage in its current state. In no way is that AI technology in the same ballpark as a human.

To illustrate the point, lets use an example of a 1st and 10 yard play about to happen in a game. AI data is smart enough to run every data point it has and spit out a likelihood of what play is coming, to which player and possible outcome. It can analyze situation, coach/player tendencies, matchups, field position, score and can do it all pretty quick.

Now the play has happened and it was a 10 yard out reception by the WR. That's about all you're going to get from the image AI. Maybe a "Number 10 throws the ball to Number 80" because that was all the data it could process in 20 seconds.
Meanwhile a human can watch the same play and diagnose what made the play successful. Who had a good block, who missed a block or an assignment, how the QB looked off the DB or the WR sold a fake and do that in a quick timeframe.

So going back to my first comment in this thread, I think AI can take over a lot of who do I draft?, who do I start? type of decisions. The FF community is still going to have to be the ones evaluating the players to determine if the stats were meaningful or a fluke or which upcoming rookies will transition to NFL better than others. Player X drops 10 pounds in the offseason and looks sharper coming out of cuts than ever before... that's a real data point that we see with our eyes, but AI will struggle to put numbers and values to that info.

Bottom line is maybe the meat and potatoes of fantasy football go the AI route, but the eye ball tests, nuances of the games still have a long time in the human domain.
 
Last edited:
AI will do everything in the next few years. So yes.

In my opinion the sharks will be shifting through the deluge of data that we'll be able to throw at AI to find what data points are relevant and what aren't, while everyone else is caught up in what is most reshared.
Won't the AI figure out which data points are relevant and which aren't?
I'm not sure the sample size with a 17 game season is enough, especially when one in game coin flip decision can impact the rest of the game. Then you throw in new wrinkles like the QB rugby scrum or an NFL rules change and the models have to be reworked with limit samples once again. Lastly, you have coaches doing dumb things that I'm not sure AI can account for.
if AI can’t find it, how are people?
1. They don't. FF is largely luck.
2. Every time someone around here says "eyeball" or "smell" test they are usually doing analysis that's not data driven.
 
Also I don't consider adding a feature to the draft dominator where I asked it who to draft and it verbally tells me who to pick AI. IMO that's a programming feature.
 
The clip below is related to music and AI towards the end. The relevant point is that for AI to gain traction is will still need a personality to go with it.

 
Yeah, because Watson does such a good job.

This is overlooked by many.

We've had AI predictive functionality with ESPN Watson's for a long time. This is not new. ESPN's projector, Mike Clay isn't going away.

What WILL happen is Mike will continue to use AI as a tool to help him predict. The same way he uses spreadsheets to crunch data and Grammarly to help him write.

I believe this:

AI isn't going to take your job.

But someone who's good at using AI tools may take your job.
 
The other thing is, the fantasy podcasts I listen to are like 50% for the info and 50% for their personality. There's lots of people out there with similar info. It's about delivery.
If a personality has alot of videos and audios for AI training, there is a potential for unlimited podcasts using technology recently created for influencers. An example, Caryn Marjorie with the help of Forever Companions. Imagine the potential for those who can't get enough of Simmons or Le Betard.
 
The other thing is, the fantasy podcasts I listen to are like 50% for the info and 50% for their personality. There's lots of people out there with similar info. It's about delivery.
If a personality has alot of videos and audios for AI training, there is a potential for unlimited podcasts using technology recently created for influencers. An example, Caryn Marjorie with the help of Forever Companions. Imagine the potential for those who can't get enough of Simmons or Le Betard.
If that's the case, it will water down the market so much I probably wouldn't bother. What is the point if we can just create an infinite number of Bill Simmons podcasts or if we can make a new Drake album every week?
 
If that's the case, it will water down the market so much I probably wouldn't bother. What is the point if we can just create an infinite number of Bill Simmons podcasts or if we can make a new Drake album every week?
There has to be audio and video for training in the first place, so there is a need for humans. At this point in time. Can AI create a new and popular personality? Or a new Drake album on demand? Perhaps. Ownership is an issue.

Most of the FF questions asked on this board could be replaced by AI. Does ownership becomes an issue if AI is using PFF grades? There's a lot of luck in FF, especially drafting. But a small edge was all those MIT students needed to win lots of $$$ in blackjack. The issue I see is that the training set for AI constantly changes year to year, and week to week. Maybe AI can be helpful with a strategy for those high-stakes FF contests. Joe won't know if we used advice from AI or from one of our resident experts
 
AI gets 1 entry in the FBG subscriber contest just like each of us subscribers. Advertise as a chance to prove you can beat “Watson.” Please make this happen! Win or lose, would be fun to analyze and track who survives longer.
 

Users who are viewing this thread

Top