‘Manuscript Writer’ AI Uses Machine Learning To Draft Science Papers

[Rengar]

Could Manuscript Writer software lead to plagiarism and non-transparency?
“Manuscript Writer”, a newly released piece of software released by sciNote claims it can save scientists and researchers time, by eliminating the tedium and boredom and the tedium of pulling together research data, findings and methodologies by using artificial intelligence to draft papers for them.

SciNote, the company behind the new software state that the average amount of time devoted to most scientific papers is some 72 hours. Manuscript Writer can cut that time down to a mere number minutes, if the company’s claims prove to be correct.

sciNote IA seems innocent enough…

Can it do my homework for me?
No. Well, Maybe. It’s not really aimed at students. And it’s not like Manuscript Writer can deliver the final polished piece of work either. Instead, SciNote says that the software gives its users a readable draft that includes “the introduction, materials & methods, results and references of your manuscript”. It’s a big claim, but one that could change the face of the science and research paper game; if it works.

Klemen Zupancic, PhD, CEO of sciNote LLC said: “While the competition within the scientific community to publish articles in high-ranking journals is constantly on the rise, it is also vital that valuable research data are published, and therefore accessible, at the earliest possible time.

“sciNote’s ELN is already used by over 20,000 scientists to store and manage scientific data. The announcement of this new AI add-on has the potential to transform the article writing process and empower these scientists, while establishing sciNote as a leader in the industry.”

Brilliant, no?
Yes. No. Depends who you speak to. Not everyone agrees. Critics have speculated that SciNote’s ‘wunder’ AI may not be as advanced as its creators suggest, and could also find itself easily susceptible to plagiarism.

Manuscript Writer’s main selling point however is that it could dramatically simplify the process and time to prepare scientific manuscripts by using current advances in machine learning and AI. Essentially, the more scientists who use it, the better it should get. But the proof will be measured in whether the science community embrace the tool or not. The company is inviting scientists to check out the software and try it fr themselves. This is the link here. One imagines that while it probably helps if you are involved in the science world, but not necessarily.

 

[Deleted member 65228]

Where does the data come from? If they are crawling the web for their Artificial Intelligence to create the drafts using data found on other webpages then it isn't a potential scenario of plagiarism, it's verified plagiarism, as long as any of the websites used to comprehend information about the topic for generating the reports do not permit unauthorised use of the content (copyright protection -> no consent given to the company for using the data).

The solution could be to credit all the sources if the Artificial Intelligence technology is smarter and can accurately track which information auto-written up used data from where and what-not, but they'd still be making money as a service providing this to use data by others who may have copyright protection (e.g. even if credit is given, one could argue that their work was used for a paid service and that is unfair because they didn't get a payment for their work being used which was free for people to learn from and not for re-sale -> why should another company make a profit over the creator for publicly available information?)

I like the idea and think it is creative and if it works then undoubtedly it can be amazing but I don't think freely accessible information posted by people should be taken and used in a paid service without credits given or the creator getting paid if it was their content which was used for the modified write-up report. It would be like a scanner to auto-generate code using snippets from StackOverflow and then someone compiling it if it really worked and selling the product, even though code which was under strict licensing to enforce open-source requirements/no commercial selling -> that wouldn't be acceptable, would it? pretty much the same scenario as this but for text based documentation instead of code.