Every March, millions of people confidently fill out their NCAA tournament brackets armed with advanced stats, historical trends and strong opinions about teams they've watched all season.
I can't do that. I'm not a college basketball expert. I follow the NBA.
Ask me about trade rumors or whether the Lakers should blow up the roster, and I'll talk your ear off. Ask me about a potential second-round matchup between a No. 4 seed from the Big Ten and a random dark-horse team, and I'm nowhere.
Last year, instead of pretending otherwise, I tried something different: I asked ChatGPT to help me fill out my March Madness bracket. The idea was simple. If millions of people were already using AI tools to summarize research, analyze data and make everyday decisions, why not see if it could help someone like me, a casual college basketball observer, make smarter picks during the most chaotic tournament in all of sports?
And the results were pretty surprising. ChatGPT hit on upsets and the collapse of favorites, and I kept climbing the leaderboard in my pool. By the time the Final Four rolled around, I was suddenly in striking distance of winning the whole thing.
Ultimately, I didn't win. But I came close enough that it didn't feel like the original joke experiment I intended, and it instead felt like something worth testing again.
So yeah, this year, I ran it back.
When the 2026 NCAA tournament bracket dropped on Selection Sunday, I once again turned to ChatGPT to build my March Madness bracket from scratch. I wasn't suddenly going to become a college basketball sicko overnight, and that's kinda the point of the experiment.
Could AI help someone who mostly watches the NBA put together a bracket that felt informed, stayed alive through the early rounds and maybe even held up deep into the tournament?
Before we get to how it all turned out, here's how ChatGPT made its picks.
What AI gets right (and wrong) about March Madness
One thing AI quickly understands about March Madness is that predicting it perfectly is basically impossible. Even if you knew the exact probability of every game outcome, the chances of picking a perfect bracket are astronomically small, 1 in 120.2 billion, and that's if you know ball. You're far more likely to be attacked by a shark.
That randomness of this tournament is exactly why it's so fun and why office pools exist in the first place.
What ChatGPT tends to get right is structure. It understands the historical patterns of the tournament. Higher seeds usually advance, certain midseeds are common upset candidates, and picking too many underdogs early is usually a mistake.
Where AI struggles is the same place most humans do: the unpredictable chaos that defines March Madness. Hot shooting nights, injuries, coaching adjustments and momentum swings all play a part.
But that's also why the experiment is interesting to me. If AI can build a bracket that survives the early rounds and avoids obvious mistakes, it might do better than the average bracket filled with random guesses in most pools.
An OpenAI spokesperson reached out to me about my story and offered this statement: "ChatGPT can be a fun and useful tool for anyone filling out their brackets this year. Whether you follow the sport closely or just want a little help making your picks, it's free to use and can break down team stats, compare matchups, and help you think through different approaches depending on how you want to play it."
How I asked ChatGPT to pick my bracket
My experiment lat year worked best when I didn't simply ask ChatGPT: "Who will win March Madness?" Instead, I gave the AI a structured prompt that tried to mimic how analysts evaluate similar tournament matchups. The goal isn't perfection but instead building a bracket that balances favorites with a few realistic upsets. The kind of bracket that might actually win a pool.
The prompt I decided on for this year:
"You are helping me fill out a March Madness bracket. Use historical tournament trends, team seeding and general basketball analytics to suggest winners for each matchup. Avoid unrealistic brackets with too many early upsets, but include a few plausible ones, including possible dark-horse and Cinderella teams. The goal is to build a bracket that could realistically win a pool. Use this as a guide: https://www.ncaa.com/march-madness-live/bracket"
Note: I used ChatGPT 5.4 Thinking.
You can adjust the prompt depending on how risky you want the bracket to be. If you're in a large pool with thousands of people, it makes sense to tell ChatGPT to be more aggressive with upset picks in order to stand out. In a smaller pool, a more conservative strategy, leaning toward higher seeds and avoiding early-round chaos, often gives you a better chance of staying competitive.
This is some info ChatGPT provided about its upset picks.
How well did ChatGPT fare?
By the end of the men's NCAA tournament, ChatGPT went 50 for 67 on my bracket, including the First Four, which is good for about 75% accuracy overall. If you remove the play-in games, that's 48 correct picks out of the main 63-game bracket, or roughly 76%. For March Madness, where even the best brackets get trampled in a matter of days, that's solid.
Still, solid isn't the same thing as winning a pool. ChatGPT did well enough in the early rounds to avoid me any embarrassment, but it missed where it mattered most -- late in the game. It projected Duke over Florida and Arizona over Michigan in the Final Four, only for the title game to end up as Michigan against UConn instead. In other words, the bracket held up through the early chaos, but it didn't stay sharp deep into the tournament.
First Four
(11) Texas over (11) NC State ✅
(16) Howard over (16) UMBC ✅
(11) SMU over (11) Miami (OH) ❌
(16) Lehigh over (16) Prairie View ❌
East Region
Round of 64
(1) Duke over (16) Siena ✅
(9) TCU over (8) Ohio State ✅
(5) St. John's over (12) Northern Iowa ✅
(4) Kansas over (13) California Baptist ✅
(6) Louisville over (11) South Florida ✅
(3) Michigan State over (14) North Dakota State ✅
(7) UCLA over (10) UCF ✅
(2) UConn over (15) Furman ✅
Round of 32
(1) Duke over (9) TCU ✅
(5) St. John's over (4) Kansas ✅
(3) Michigan State over (6) Louisville ✅
(2) UConn over (7) UCLA ✅
Sweet 16
(1) Duke over (5) St. John's ✅
(2) UConn over (3) Michigan State ✅
Elite Eight
(1) Duke over (2) UConn ❌
West Region
Round of 64
(1) Arizona over (16) LIU ✅
(9) Utah State over (8) Villanova ✅
(5) Wisconsin over (12) High Point ❌
(4) Arkansas over (13) Hawaii ✅
(6) BYU over (11) Texas ❌
(3) Gonzaga over (14) Kennesaw State ✅
(7) Miami (FL) over (10) Missouri ✅
(2) Purdue over (15) Queens ✅
Round of 32
(1) Arizona over (9) Utah State ✅
(4) Arkansas over (5) Wisconsin ✅
(3) Gonzaga over (6) BYU ✅
(2) Purdue over (7) Miami (FL) ✅
Sweet 16
(1) Arizona over (4) Arkansas ✅
(2) Purdue over (3) Gonzaga ✅
Elite Eight
(1) Arizona over (2) Purdue ✅
South Region
Round of 64
(1) Florida over (16) Lehigh ✅
(8) Clemson over (9) Iowa ❌
(5) Vanderbilt over (12) McNeese ✅
(4) Nebraska over (13) Troy ✅
(6) North Carolina over (11) VCU ❌
(3) Illinois over (14) Penn ✅
(7) Saint Mary's over (10) Texas A&M ❌
(2) Houston over (15) Idaho ✅
Round of 32
(1) Florida over (8) Clemson ✅
(4) Nebraska over (5) Vanderbilt ✅
(3) Illinois over (6) North Carolina ✅
(2) Houston over (7) Saint Mary's ✅
Sweet 16
(1) Florida over (4) Nebraska ✅
(2) Houston over (3) Illinois ❌
Elite Eight
(1) Florida over (2) Houston ❌
Midwest Region
Round of 64
(1) Michigan over (16) Howard ✅
(8) Georgia over (9) Saint Louis ❌
(5) Texas Tech over (12) Akron ✅
(4) Alabama over (13) Hofstra ✅
(6) Tennessee over (11) SMU ✅
(3) Virginia over (14) Wright State ✅
(10) Santa Clara over (7) Kentucky ❌
(2) Iowa State over (15) Tennessee State ✅
Round of 32
(1) Michigan over (8) Georgia ✅
(4) Alabama over (5) Texas Tech ✅
(3) Virginia over (6) Tennessee ❌
(2) Iowa State over (10) Santa Clara ✅
Sweet 16
(1) Michigan over (4) Alabama ✅
(2) Iowa State over (3) Virginia ❌
Elite Eight
(1) Michigan over (2) Iowa State ✅
Final Four
(1) Duke over (1) Florida ❌
(1) Arizona over (1) Michigan ❌
National Championship
(1) Duke over (1) Arizona ❌