Article on The Atlantic about Grammarly and AI

It may be behind a paywall (I subscribe), but someone can normally find an accessible version:


What Was Grammarly Thinking?

A short-lived AI tool promised to help users write like the greats—and a bunch of other random people, including me.
By Kaitlyn Tiffany
A precis is that Grammarly added an ‘Expert Review’ feature which allowed you to get suggestions for improvement ‘inspired by [insert author or journalist here].’

None of the authors or journalists we’re asked for permission to use their names. One is suing Grammarly’s parent organization. And the ‘advice’ to use a technical term, was utter crap.

Grammarly’s chat-bot even denied that the company was doing what it was doing. The CEO posted an apology on LinkedIn.
 
They have agreed to shut down the "feature" now. But there are multiple law suits.

From my understanding, they are pretty much dead meat in California where it is illegal to appropriate someone's identity for commercial purposes. (@SimonDoom might be able to clarify if that is really the case, but that is how I saw it explained.)
 
How dare they not waiting 100 years after their death to revive them with their technocratic necromancy 2nd-level spells, now they have to roll a 1 on Stealth twice after spending that Inspiration!
 
There is a second article on The Atlantic today which describes multiple tech bros saying that bubbles are good for you. Sure they might crush the economy and - you know - actual people, but at least the AI firms can raise capital. So… er… yay!

Even Silicon Valley Says that AI Is a Bubble

An AI crash could bring down the economy. Some in the tech world think that's the price of progress.
By Lila Shroff
 
They have agreed to shut down the "feature" now. But there are multiple law suits.

From my understanding, they are pretty much dead meat in California where it is illegal to appropriate someone's identity for commercial purposes. (@SimonDoom might be able to clarify if that is really the case, but that is how I saw it explained.)

Yes. Civil Code section 3344, which prevents the unauthorized use of another's name, likeness, voice, signature, or photograph for commercial purposes. I haven't heard of its use in this specific context but it seems like it would apply.

Regardless of the law, what Grammarly did is just shitty. It's going to be interesting to see if legislatures respond by trying to regulate this stuff, or if the AI tycoons are so powerful that they shut the regulatory attempts down.
 
To think I missed all the news about Grammarly. No wonder it seems to shake and sigh now and again as I write. Should ‘shirely’ be ashamed of themselves...
 
Yes. Civil Code section 3344, which prevents the unauthorized use of another's name, likeness, voice, signature, or photograph for commercial purposes. I haven't heard of its use in this specific context but it seems like it would apply.
Interesting argument. Does the word 'voice' include 'literary voice'? I'm glad someone's clarifying this law at their own expense, and not mine.
 
Interesting argument. Does the word 'voice' include 'literary voice'? I'm glad someone's clarifying this law at their own expense, and not mine.

I'm not aware of any precedents that say that. I could imagine an argument against someone who compiles Author X's work, plugs it into an AI tool, and then says "Here's how Author X would change your story" and tries to make money from doing that by using Author X's name.
 
I'm not aware of any precedents that say that. I could imagine an argument against someone who compiles Author X's work, plugs it into an AI tool, and then says "Here's how Author X would change your story" and tries to make money from doing that by using Author X's name.
I can imagine many interesting arguments, this way and that. I'm a lawyer.

Now, you can imagine some Lit author asking AI to render his story in the authorial voice of 'Famous living author'(FLA) and publishing it on Lit for free, he's a recreational writer. Is there no 'commercial purpose'? Can Lit's 'commercial purpose' be attributed to the recreational Lit author. Can the Lit authors use of FLA's name in their prompt be imputed to Lit? (No - Lit has immunity, but imagine it hasn't) We need to know, because that's the future, and not all hosts have immunity.
 
Last edited:
There aren’t they are two different articles, written by two different journalists, both from the same site.
 
Last edited:
It also seems to imply that they are illegally training it on published works of living authors. I mean, thats the only way it would have worked, if it even had.
 
It also seems to imply that they are illegally training it on published works of living authors. I mean, thats the only way it would have worked, if it even had.
That's one of the points of interest. Is the training illegal? Some people claim it is, but there's no ruling to that effect. Some trainers have offered compensation schemes in class actions 'without prejudice'. Don't forget to add your name to the class action suits to get your share of the fund.

AI Experts are all the rage at the moment; everyone's producing them. They rely entirely on limiting their domain of expertise to authoritative sources of fact and opinion. No consent is sought from those whose writings contribute to the domain. The Polytechnic University of Hong Kong Faculty of Law has just introduced an AI Agent which will be a formal component of its instruction. DeepSeek boasts over 380 AI Experts. If I had to bet, I'd bet on this sort of usage being a transformational, fair usage.

Grammarly's promotion of its Expert feature may have been clumsy, in that it openly flaunted the source of its expert opinion instead of saying 'Ask a writer, or writers, of your choice who advise on writing, how they would improve a passage of your writing'. Will that make any difference to the outcome.? Will Grammarly offer the aggrieved writer/s a few bob to go away?
 
Back
Top