Do you really think that the LLM knows or cares whether you get fired? Your mistake here is treating LLMs are sentient when all they really are is a token probability predictor.

Why do users insist on using tools that can get them fired? This is either a policy or culture problem, not an AI problem.

Hell, just make a macro to replace all the funny anachronisms and a prompt that can reference your own writing style to massage the output.

“This isn’t written by Ai”

‘And neither is this’

My iPhone and my computer do it automatically.

How about not using a LLM if you could get fired for doing so...

Quotes like that were an issue long before AI. Classically, it was a sign that something was copied from Word or Outlook. I used to run into it constantly 15 years ago, I had a co-worker run into it a couple weeks ago, no AI involved.

As much as I’ve seen this, I’ve never seen anyone get fired for it, this includes times where it broke production.

There are many more places these days where not using AI gets you fired… :)

I think some programs on some computers will automatically use curved quotations marks (even if they can be disabled, it is not always done), and sometimes they are deliberately entered manually, so that is not necessarily the AI.

Using curved, matching quotation marks is typographic design and style, as is using hyphens, en, and em dashes correctly in context. Many tools will automatically use the correct typographic symbols in context in the same way they’ll automatically insert periods and spaces in response to double-spacing. And even setting that aside—typographic style is a long-standing craft of design. Why would we want to let LLMs take that away from us?