5 reasons why including ChatGPT in your list of authors is a bad idea

https://garymarcus.substack.com/p/scientists-please-dont-let-your-chatbots

Scientists, please don’t let your chatbots grow up to be co-authors

Five reasons why including ChatGPT in your list of authors is a bad idea

The Road to AI We Can Trust

@garymarcus I was part of a discussion on past and engineering computer engineering tools yesterday on this website and you know what the dawn of #ChatGPT reminds me of?

The dawn of (accessible) FEA (finite element analysis) in mechanical engineering.

There was a time when they first started putting #FEA packages in #CAD packages that novice engineers would run a simulation, produce a "pretty picture" and just trust the results without scrutiny.

That brings me back.

@garymarcus this bot is just like some who sounds scientific but actually know nothing about science.
@charleston @garymarcus I've seen it called "Mainsplaining as a service"
@garymarcus At this point the legit use is mundane/routine things you can readily verify, test.

@garymarcus

Usually we discuss authorship related topics with our co-authors - how about we ask ChatGPT if it wants to be listed?

Sounds like it has a very clear opinion on its contribution...

@VisionBernie @garymarcus your question seems to rely on the assumption that it's a potential co-author, which imho ascribes too many anthropomorphic qualities to it
@garymarcus "Friends don’t let friends co-author with ChatGPT" - thank you
@garymarcus @DrVeronikaCH Same for GPT3, right? See this paper who tests this and makes a distinction between ChatGPT and GPT3 https://www.linkedin.com/posts/almiraosmanovicthunstrom_gpt3-activity-7015410303953698816-jayN
Almira Osmanovic Thunström on LinkedIn: GPT3 | 14 comments

When we submitted the paper which GPT-3 wrote on itself last year we looked forward to the reviewer comments. The paper went on a WILD ride in the academic… | 14 comments on LinkedIn

@cgruenloh @garymarcus I'm happy for these tools to be acknowledged in the same way as Grammarly would be
@cgruenloh @DrVeronikaCH same for any LLM thus far or in near-term future!
@garymarcus Great post! I made bladerunner.ai to highlight GPT text so that readers can fact check
@garymarcus Good points. Also, by the rules of good scientific practive, all co-authors share responsibility for the paper. As a software tool ChatGPT cannot take responsibility.