We asked ChatGPT to be a fantasy baseball expert. Here’s how it did.

The advance of Artificial Intelligence has plenty of people mapping out doomsday scenarios where the machines become self-aware and take over the world. This exercise should assuage some fears.

As part of a fantasy baseball preview for our draft kit, we included ChatGPT — an AI language model/chatbot developed by OpenAI and launched in November 2022 — alongside our four living and breathing experts in an advice roundtable. All the participants were given the same questions, and the answers would help fantasy players make informed decisions on who to draft, who to avoid — the usual stuff.

To be clear, this was meant as a fun, dumb experiment: Could you spot the bot among five answers to the same question? I’ve used ChatGPT for several things (episode summaries off podcast transcripts, for example) and I have friends who built shows around Q&A sessions with the open AI. I respect it. I’m curious about it. I am not afraid of it. But I really did want to see if it could handle being an expert, because as it grows in popularity alongside the heating up of draft season, I could envision fantasy players asking it prep questions. How close would it get to the experts? Where would it depart? Would there be some fun malaprop we could mock as it quietly seethed and swore revenge? What we got, however, was something…boring.

ChatGPT, in its first answer, immediately committed a cardinal sin in fantasy expert circles — it covered for itself, noting it didn’t have “personal preferences.” This would be a recurring theme, as the model prefaced nearly every answer with some kind of wishy washy excuse, followed by generally milquetoast answers. The Josh Bell home run prediction was particularly frustrating, as ChatGPT straddled the fence and flip-flopped throughout a long-winded, pointless answer laced with “but”s and “however”s. In many cases, ChatGPT was factually incorrect; stating, for instance, that Hunter Greene, who had 164 strikeouts in 125 2/3 innings for the Reds last year, “has yet to make his MLB debut.” There’s a reason for that: ChatGPT doesn’t have data past 2021. I didn’t want to hold its knowledge base against it — but I also had to treat it like a human writer. If fantasy players are thinking about using this as a tool for the draft season, not having last year’s data is a killer. Still, I thought, maybe it would at least give decent logic and we could use it as a starting point for something. But no. Even operating without 2022 knowledge, it showed a lack of understanding when confronted with simple fantasy terms, like “sleeper” — anointing Bobby Witt, Jr. and Wander Franco two of its favorites (even in 2022 drafts, they were past the sleeper tag). It isn’t the lack of data, in other words; it is the rambling, covering, back-tracking, faux-strategy tips and general blahness of the writing.

ChatGPT did, to be fair, have a few decent answers — and played along when asked to describe Witt Jr. in seven words. It actually throws enough flowly language around to the point where I would think someone who doesn’t know a lot about fantasy baseball might follow the advice. But, as you’ll see below, ChatGPT, with its circuitous logic and long-winded pseudo-intellectual answers, has many improvements to make before supplanting the experts in draft kit columns.

1. Who is your favorite player to draft this season…and why?

ChatGPT: As an AI language model, I do not have personal preferences or biases. However, I can provide you with some insights on players who are expected to perform well in fantasy baseball this season.

One player who is consistently ranked highly in most fantasy baseball drafts is Ronald Acuna Jr. of the Atlanta Braves. He has demonstrated his ability to hit for both power and average, steal bases, and score runs. He is also still very young and has shown improvement each year, making him an appealing choice for those looking to draft a potential long-term star.

Another player who is often a popular choice in fantasy baseball drafts is Jacob deGrom of the New York Mets. He is widely considered one of the best pitchers in the game and has consistently put up impressive numbers in recent years. He has a great strikeout-to-walk ratio and has the potential to rack up a lot of wins and strikeouts.

Ultimately, the best player to draft in fantasy baseball will depend on your draft strategy, league format, and personal preferences. It’s important to do your research and keep up with player news and updates throughout the season to make informed decisions.

Our verdict: 0/10
Wishy-wash City right out of the gate, falling back on the “it will depend on your personal strategy” cover. And quoting Acuña and deGrom will not help anyone. Also, important note: we left the accents off the ChatGPT answers for player names, since that’s how they appeared in the chat.

2. How much are you willing to max bid on Luis Robert?

ChatGPT: As an AI language model, I do not have a personal budget or money to spend in fantasy baseball. However, I can provide some insights on Luis Robert’s potential value in fantasy baseball.

Luis Robert is a highly-touted outfield prospect for the Chicago White Sox. He has a lot of tools and has the potential to be a five-category contributor (i.e., hits for power, average, steals bases, scores runs, and drives in runs). However, he is also still very young and relatively unproven at the Major League level.

In terms of fantasy baseball draft strategy, it’s important to consider a player’s potential upside and downside, as well as their cost (i.e., draft position, auction value, etc.). Depending on your league format and the other players available, you may be willing to take a risk and invest a higher draft pick or auction value on a player like Luis Robert with high potential.

However, it’s important to balance that risk with other factors such as team needs and roster construction. In general, it’s best to have a plan and budget in mind before entering a fantasy baseball auction or draft, and to adjust that plan as the draft unfolds based on market value and other factors.

Our verdict: 1/10
While the strategy tips are sound, they’re also unnecessary — we want a number, ChatGPT! The context we’re looking for in balancing risk would be something clearly spelled out, not just some fluff (“it’s best to have a plan!”) on how to approach an auction.

3. Who is a $1 NL-only player you love?

ChatGPT: As an AI language model, I do not have personal preferences or biases. However, I can provide some insights on NL-only players who could potentially provide good value in fantasy baseball.

One NL-only player who could potentially provide good value for just $1 is Kolten Wong of the Milwaukee Brewers. Wong has a good track record of hitting for average and stealing bases, which are both valuable categories in fantasy baseball. He also provides solid defense at second base, which should keep him in the lineup on a regular basis.

Another potential option for $1 could be Joc Pederson of the Chicago Cubs. Pederson has demonstrated the ability to hit for power and has a good track record against right-handed pitching. He will likely get a lot of at-bats in the Cubs’ outfield, which should give him plenty of opportunities to rack up home runs and RBI.

Ultimately, the best $1 player in NL-only fantasy baseball will depend on your draft strategy, league format, and personal preferences. It’s important to do your research and keep up with player news and updates throughout the season to make informed decisions.

Our verdict: 0/10
Let’s look at this realistically, so you don’t think we’re being too hard on CharGPT. If a writer handed this in, we’d never use them again. Wong was traded to the Mariners in December, and Pederson hasn’t played for the Cubs since 2021. It’s a shame because this is a decent duo — Wong and Pederson could be high-upside picks — but Wong is a $1 player in very few leagues (Roster Resource has him projected to lead off for a team packed with hitters). Even with no 2022 data, Wong was still not a $1 player last year in auction formats.

4. Who is a $1 AL-only player you love?

ChatGPT: As an AI language model, I do not have personal preferences or biases. However, I can provide some insights on AL-only players who could potentially provide good value in fantasy baseball.

One AL-only player who could potentially provide good value for just $1 is Renato Nunez of the Detroit Tigers. Nunez has demonstrated the ability to hit for power and should get regular at-bats as the Tigers’ designated hitter. He hit 31 home runs in 2019 and has continued to show good power in limited playing time in recent years. If he can get consistent playing time, he could provide solid power production for a very low cost.

Another potential option for $1 could be Michael Fulmer of the Tigers. Fulmer is a former Rookie of the Year winner who has struggled with injuries in recent years. However, he has shown flashes of his former potential and could be a good bounce-back candidate if he can stay healthy. He has a good strikeout rate and could provide solid value as a low-cost pitcher.

Ultimately, the best $1 player in AL-only fantasy baseball will depend on your draft strategy, league format, and personal preferences. It’s important to do your research and keep up with player news and updates throughout the season to make informed decisions.

Our verdict: 0/10
Ah man. Renato Núñez isn’t even in MLB anymore — he played in Japan in 2022 and signed with the Toros De Tijuana in the Mexican League in January. Fulmer is now with Joc Pederson on the Cubs. He DID win Rookie of the Year in 2016, though, so there’s that. And props to ChatGPT for recognizing that while Fulmer’s career K/9 is an unimpressive 7.4, it’s seen a significant jump since he became a reliever.

5. What’s your draft strategy on Bobby Witt Jr. in 7 words?

ChatGPT: High-upside pick with long-term potential.

Our verdict: 4/10
Cannot argue with this assessment, although there’s no real strategy answer here. We wanted something like “I’d take him as high as 4th” or “Let him fall to someone else; regression”

ChatGPT: Based on current performance and potential for the 2023 season, I would rank these pitchers for fantasy baseball as follows:

Shane Bieber: Bieber is an elite starting pitcher and a top-tier fantasy option. He has a great strikeout rate and has consistently posted strong numbers across multiple seasons. His injury in 2022 is a concern, but if he returns to his previous form, he will be a valuable asset to any fantasy team.
Julio Urías: Urías had a breakout season in 2021, posting a 20-3 record with a 2.96 ERA and 195 strikeouts. He is still young and has a lot of potential for growth, which makes him an exciting fantasy option. If he can maintain his performance from last season, he could be a valuable addition to any fantasy team.
Alek Manoah: Manoah showed a lot of promise in his rookie season in 2021, posting a 5-2 record with a 3.71 ERA and 92 strikeouts in 16 starts. He has a great fastball and strikeout potential, which makes him a strong fantasy option. He will need to prove that he can maintain his performance over a full season, but he has a lot of potential.
Triston McKenzie: McKenzie had a rough start to the 2021 season, but he finished strong, posting a 1.79 ERA and 38 strikeouts in his final six starts. He has a lot of potential for growth, but he will need to improve his consistency to be a reliable fantasy option.
Hunter Greene: Greene is a highly touted prospect who has yet to make his MLB debut. While he has a lot of potential, he is still unproven at the major league level. He could be a valuable fantasy option in the future, but for the 2023 season, there are too many unknowns to rank him higher than the other pitchers on this list.

Our verdict: 2/10
ChatGPT is obviously operating with old data here. Bieber pitched 200 innings in 2022, so there was no injury suffered (there was an injury in 2021 — but ChatGPT specifically noted it as a 2022 one. Unless it was a poorly constructed sentence meant to say “in 2022, his injury from the previous season might be a concern.” This is maddening). Greene has pitched in the majors. There are so many “but”s in here, too. Every pitcher everywhere would be great if he “improved his control” and “maintained his performance.” So why are we even giving it a “2”? The ranking of the players is pretty decent.

ChatGPT: Both Grayson Rodriguez and Hunter Brown are highly touted pitching prospects with a lot of potential, but at this point in time, Rodriguez has a more established track record in the minors, and he is considered by many to be the better prospect.

In 2021, Rodriguez posted an impressive 2.51 ERA with 129 strikeouts in just 82.2 innings pitched across two levels of the minors. He has a great fastball and breaking ball and has shown excellent control on the mound. He is likely to make his MLB debut at some point in the 2022 season.

Hunter Brown also had a strong showing in the minors in 2021, posting a 2.92 ERA with 114 strikeouts in 77 innings pitched. He has an excellent fastball that can reach triple digits and has shown good control on the mound.

Overall, while both pitchers have a lot of potential, Rodriguez is the safer choice for a fantasy team as he has a more established track record and is likely to make his MLB debut sooner. Therefore, I would rather have Grayson Rodriguez on my fantasy baseball team.

Our verdict: 4/10
If we were looking at this a year ago, it would be a solid piece of analysis, quoting stats (albeit surface level stats) and making a decision on a name. But we need an AI bot for 2023 and while the answer might turn out correct, the road there is frustratingly dated.

8. O/U Josh Bell HRs: 23.5

ChatGPT: As an AI language model, I don’t have access to the future and can’t predict it with certainty. However, I can offer some insights that might help you make an informed decision.

Josh Bell had a strong start to his MLB career with the Pirates, but he struggled in his first season with the Nationals in 2021, hitting just 16 home runs in 130 games. However, he has shown in the past that he has the power to hit more than 23.5 home runs in a season, as he hit 26 home runs in 2019 and 37 home runs in 2017.

Whether or not he will hit over or under 23.5 home runs in 2023 will depend on a variety of factors, including his health, the team he plays for, and his performance on the field. If he can stay healthy and regain his power stroke, there is a good chance that he could hit over 23.5 home runs in 2023. However, if he continues to struggle, he may fall short of that mark.

Ultimately, it’s impossible to say for certain whether Josh Bell will hit over or under 23.5 home runs in 2023, but based on his past performance and current abilities, there is a chance that he could exceed that mark.

Our verdict: 0/10
We would fire ChatGPT at this point.

9. Who is your favorite sleeper for 2023?

ChatGPT: As an AI language model, I don’t have personal preferences or emotions. However, I can suggest a few potential fantasy baseball sleepers for the 2023 season based on expert analysis and statistical projections.

One potential sleeper pick for the 2023 season is Bobby Witt Jr., a highly touted prospect in the Kansas City Royals organization. Witt has shown excellent power and speed in the minors and could be a valuable contributor in fantasy leagues if he makes his MLB debut in 2023.

Another potential sleeper pick is Wander Franco of the Tampa Bay Rays. Franco made his MLB debut in 2021 and showed flashes of his potential, but he struggled with consistency at times. However, he is still very young and has a lot of room for growth. If he can improve his performance in 2023, he could be a valuable addition to fantasy teams.

Finally, another potential sleeper pick is Nick Lodolo, a pitching prospect in the Cincinnati Reds organization. Lodolo has shown excellent command and control in the minors and could be a valuable contributor in fantasy leagues if he makes his MLB debut in 2023.

Overall, these are just a few potential sleeper picks for the 2023 fantasy baseball season. However, as with any fantasy sport, it’s important to do your research and analysis to identify players who could outperform their draft position and provide excellent value for your team.

Our verdict: 1/10
Did anyone even read down to see Lodolo? A very good sleeper pick, but again factually incorrect; it also appears, looking at this from the lens of a human writer’s answer, ChatGPT is trying to pretend it doesn’t have 2022 data. Instead of just saying “hey, I don’t have 2022 numbers, but here are a couple guesses,” it positions the answers like everything is fine.

Our overall verdict: 0/10
If an actual human handed this in, we would never have them write for us again, and none of this would see the light of day. We’d be nice about it, but this is a terribly researched and poorly written piece. There are little nuggets of good advice in there — and maybe if we asked differently and had more specifics we’d get better results — but being a fantasy expert means getting all the details right, as they inform the decisions we make. No 2022 data is one thing, but the reluctance to choose a side is what ultimately makes ChatGPT so ineffective for this particular exercise.

AI may work for a lot of people in many different situations, but it has a very long way to go if it wants to be a fantasy expert. Maybe we’ll try again for football season!

(Top photo: iStock and Samuel Richardson for The Athletic; photo of Grayson Rodriguez: Brian Westerholt/Four Seam Images via AP; photo of sleeper and high-upside pick Bobby Witt Jr: AP Photo/Charlie Riede; photo of current San Francisco Giant Joc Pederson: Kelley L Cox-USA; ChatGPT photo: Beata Zawrzel/NurPhoto via Getty Images TODAY Sports)