ChatGPT is fun, but not an author(science.org)
science.org
ChatGPT is fun, but not an author
https://www.science.org/doi/10.1126/science.adg7879
7 comments
Not really. The actual text of the Editorial Policy recites:
> Text generated from AI, machine learning, or similar algorithmic tools cannot be used in papers published in Science journals
which you may read in light of
> artificial intelligence tools cannot be authors
and you may complete it with the text in the article
> Machines play an important role, but as tools for the people posing the hypotheses, designing the experiments, and making sense of the results. Ultimately the product must come from — and be expressed by — [the mind of said people]
So, the point is very easy: /the content/ of all material must be generated by the researcher, while AI can be - as always - a tool to aid production.
--
About «defining "AI tool"»: as in the past seven decades - there exist natural problem solvers, and there can be automated problem solvers. AI automates what a professional does. Edit: disambiguating: "[content] generated from algorithmic tools [with the purpose of generation]" are forbidden; "AI tools [with the purpose of being tools, aids] cannot but be welcome". The policy is badly formulated, though the sense should be clear.
> Text generated from AI, machine learning, or similar algorithmic tools cannot be used in papers published in Science journals
which you may read in light of
> artificial intelligence tools cannot be authors
and you may complete it with the text in the article
> Machines play an important role, but as tools for the people posing the hypotheses, designing the experiments, and making sense of the results. Ultimately the product must come from — and be expressed by — [the mind of said people]
So, the point is very easy: /the content/ of all material must be generated by the researcher, while AI can be - as always - a tool to aid production.
--
About «defining "AI tool"»: as in the past seven decades - there exist natural problem solvers, and there can be automated problem solvers. AI automates what a professional does. Edit: disambiguating: "[content] generated from algorithmic tools [with the purpose of generation]" are forbidden; "AI tools [with the purpose of being tools, aids] cannot but be welcome". The policy is badly formulated, though the sense should be clear.
I agree 100% with you. could it write an introduction better than me or take an outline and create a paragraph that's better written then I could probably.
> write [...] better than me
Very hopefully not, because we suppose you would check the content before and after expression.
Check it in terms of intelligence: in terms of relevance, of truth, of importance, of logic, of foundational solidity...
Very hopefully not, because we suppose you would check the content before and after expression.
Check it in terms of intelligence: in terms of relevance, of truth, of importance, of logic, of foundational solidity...
I suck at writing. I've always hated it.
Well, you are then more than just excused to use tools which help you in expressing yourself (contextually, "expressing the ideas that come from your own work") - as long, of course, as those tools do not make you even weaker.
But if you have that weakness, we should warmly recommend that you read a bit more, a bit better - more "quality" texts read in "quality" time -, and that you spend more attention in the activity of your expression... There is no need to keep that weakness, if it is just a matter of underdevelopment.
--
Back instead to the topic: "artificial deliriousness" cannot replace your intelligence (it has none), so there are no alternatives - an Intelligence duly has to check outputs...
But if you have that weakness, we should warmly recommend that you read a bit more, a bit better - more "quality" texts read in "quality" time -, and that you spend more attention in the activity of your expression... There is no need to keep that weakness, if it is just a matter of underdevelopment.
--
Back instead to the topic: "artificial deliriousness" cannot replace your intelligence (it has none), so there are no alternatives - an Intelligence duly has to check outputs...
This, to me, seems overreaching. As mentioned in the article, if the authors review and are accountable for generated text, it seems like an incremental (significant, but incremental) change from something like Grammarly.
Also, defining "AI tool" is going to be increasingly difficult.