Will we have AI that can explain chess moves logic in human language by 2030?

#AI #Technology #Technical AI Timelines #Chess

107

Ṁ1.1k

Ṁ750

2030

94%

chance

ALL

Will we have (and put in a comment below) an automated system that will explain chess moves logic in free text?

Requirements:

Will be available for free, as a service or open source project that I can run on a regular computer.
Will work with any reasonable real-life chess positions.
Will explain the motivation for a specific move, both for easy moves ("white recapture the queen") and more advanced logic ("aiming for exchanges to defend from the kingside attack")
Will not hallucinate at all in my testing.

Get Ṁ600 play money

Related questions

Will a large language model beat a super grandmaster playing chess by 2028?

42% chance

Will chess be solved by 2040?

12% chance

Will an AI by OpenAI beat a super grandmaster playing chess by 2028?

53% chance

Will AI enable a successful conversation between a human and a member of a non-human species by the end of 2030?

71% chance

Will chess be solved by 2100?

27% chance

Will AI be able to generate correct images of a chess game in 2024?

34% chance

Will AI surpass humans in conducting scientific research by 2030?

31% chance

Will AI extend English before 2031?

47% chance

Will an LLM (a GPT-like text AI) defeat the World Champion at Chess before 2035?

52% chance

Will AI create philosophy before 2030?

Sort by:

This is not fully free (They give 1 free review a day), but https://www.chess.com has a Review

feature which is very similar to what you want. I'm not sure whether it's an AI, but it does explain chess move logic.

Chess.com - Play Chess Online - Free Games

Play chess online for free on Chess.com with over 150 million members from around the world. Have fun playing with friends or challenging the computer!

@bohaska interesting. Can you share some examples for how it looks like?

@0482 Sure!

It can be better, but we're 90% of the way there already without LLMs.

@bohaska this is really cool and promising. Too vague sometimes for actually settling the question, but definitely a sign of the process

The strongest argument against this happening is that once you build a system to extract human-comprehensible information from the search tree, presenting it as just plain text seems highly suboptimal

It would probably make more sense to build an interface that combines text description with visualization's of selected lines/positions from the search tree + allows you to ask questions like "why doesn't this move work here?". Overall more of an interactive exploration/conversation based on the tree...