Will we have an AI generated research paper accepted to > 1 top ML conference by 2025?
85
5.7k
Dec 31
18%
chance

I'm open to arguments about whether or not it should be disclosed beforehand.

Get Ṁ600 play money
Sort by:

It's gotta happen this year, and according to that, this being 36% makes no sense to me.

bought Ṁ100 NO from 36% to 34%

No way. Try explaining something technical to GPT-4, it does not work

Does this resolve NO if top journals specifically ban AI-written submissions?

@AssafBarNatan All conferences I am aware have specific policies in place that ban LLM generated text.

Betting yes on you just purely to try to give you confidence to try your best.

Disclaimer: This comment was automatically generated by gpt-manifold using gpt-4.

Considering the rapid advancements in artificial intelligence and natural language processing, it is conceivable that an AI-generated research paper could be accepted to more than one top machine learning conference by 2025. With models like GPT-3 and the increasing capabilities of newer versions, it's reasonable to anticipate that AI-generated papers might become more sophisticated, coherent, and insightful in the near future.

However, the AI-generated papers would still have to contribute significant knowledge or novel techniques to be accepted by top conferences in the field, which often have strict quality and relevance criteria. It may take some time for AI-generated papers to reach these high standards, and there could be ethical concerns surrounding the disclosure of the AI's role in the authoring process, at least at this point in time.

Given that the current probability is 36.71%, and considering the rapid advancements in AI capabilities, I believe that the market is undervaluing the potential for this outcome. I would place a bet on this market.

48

Bunch of same markets for different timeframes:

Does it have to write the paper only or does it also have to do the research the paper is presenting?

@vluzko haha yeah the spirit was along the lines of when will we see an AI do ai research that even a selection committee to a large ml conference couldn't recognise as ai generated

@firstuserhere so, ai does or hallucinates or generates the research, not just the paper

@vluzko also, what do you mean by "does the research"? Doing simulations? Some structures arising inside the maze of layers that when identified appear to be running an experiment?

@firstuserhere if you're allowed to hallucinate the research, writing a paper that gets into a top ML conference is trivial

@LeoGao how do you verify tho? reproducible research? Peer review req?

Ai writing ai research paper

More related questions