Are any of you using (non-language translation) Artificial Intelligence (AI) in your publishing processes?


I have been tasked with researching and documenting the use of AI within the publishing industry, specifically within STM publishing, and know that this group has a wide variety of expertise they have shared for many years. 

Jennifer Goodman has already generously shared ways that RWS is using AI within the company. We are not interested in translating our English content into other languages. We are interested in other uses of AI. 

Are any of you using AI for any processes in publishing that aren't related to language translation? For instance, to write python, xquery, or even perl? Is anyone using tools in their editorial group to generate text or metadata such as keywords or topics, alt-text/metadata for images, summaries--or anything else for that matter.  

Are there tools you have looked at and decided not to use them? 

Thanks for any insights you may be able to offer.

Jaye Mize (she/her/hers)
Director, Content Production Systems

JAMA NetworkTm
330 N Wabash Ave, Ste 39300, Chicago, IL 60611

  • Hi Jaye.   I've been researching AI for about a year. I've used it to generate Perl, XSLT, python, XML Schemas, javascript, JSON structures.    Overall, I find AI to be helpful but not the "Swiss army knife" of code-writing the media and most people I've spoken to, seem to think it is.

    For instance, very little code suggested by chatbots - even paid-for platforms -  (and I've tried quite a few now), works without some sort of adjustment/editing.  Some code produces errors or don't actually solve the problem presented. 

    In summary, as most bots' creators point out, their output is not guaranteed 100% correct.   In other words: it still requires HUMAN intelligence to discern the valid and accurate from the vague and erroneous.

    I've not used AI for generating text but it might be more successful in that regard, but the caveat of assuming accuracy and validity still stands, in my opinion.