Although I am now retired, I'm still interested in following news and trends in the technical communication field. The biggest thing to affect the field recently has been the rise of language-generative AI tools like ChatGPT. Although they have notable weaknesses (they often lie and make things up) they have their uses, as noted by Rhonda Bracey in her CyberText Newsletter.
A recent post by Tom Johnson in his I'd Rather Be Writing blog takes a more detailed look at how these tools might affect the field. After describing some of the more interesting AI-based writing tools, he looks at both their weaknesses and their strong points and describes some of the uses that writers might find for them.
From his list, these are some of the things I would have used these tools for:
Simulating reader responses to documentation based on personasMining through hundreds of support tickets to identify trendsIdentifying the impact of the latest release on hundreds of existing documentation pages, including code samples.Comparing API documentation against a best practices template and identifying areas of inconsistency.Automatically linking related topics in the documentation for easy navigation.
Tom identifies several other potential uses, most of which didn't apply to my job, but would certainly help in different work environments. Some of them are similar to tasks for which I used
sophisticated macros provided by Paul Beverley, but the AI-based tools promise to be both more powerful and easier to use.
What worries me the most is the potential for uninformed managers to decide that these tools now guarantee that "anyone can write" and lay off experienced technical writers. I've seen the effects of that attitude in more than one company.
Update: Here's a
YouTube video from Candlelight Editing about how freelance editors can use ChatGPT. You can
download the prompt file that includes the prompts used in the video.
No comments:
Post a Comment