Is Grammarly Generative AI? If So, Do Lawyers Need to Disclose Its Use?
By now, you may have several AI tools in your toolkit. You probably have already taken advantage of AI-assisted research. Then there's ChatGPT and Google Bard. But as most lawyers know, relying wholly on generative AI is risky, at best.
One of the most popular AI tools is Grammarly, the ubiquitous spelling and grammar-checker that has become increasingly sophisticated in its recommendations. And yes, we did run this blog through Grammarly (a couple of times).
This begs the question — is Grammarly's functionality at a point where lawyers who use it need to disclose it as generative AI to the court? Do Florida lawyers need to disclose it to their clients to comply with Florida’s ethics advisory opinion on generative AI?
And what about privacy concerns? Does Grammarly use what you've entered to improve its algorithm? And could that breach client confidentiality?
College Student Suspended for Grammarly Use
While no specific court or bar has expressed an opinion on the use of a particular AI platform, just generative AI generally, recent news in higher education might imply it is only a matter of time before this question comes up.
University of North Georgia student Marley Stevens was suspended in February after a paper she submitted was flagged for AI use. According to Stevens, who spoke to several media outlets after her TikTok video about it went viral, she used the free Grammarly software that only checks for spelling and grammar issues.
The University of North Georgia has not revealed its side of the story, saying it cannot speak on the issue to protect student confidentiality. But, it has not indicated it is considering removing her suspension.
Generative AI vs. Other AI Tools
The key question here is on generative AI, programs that can be used to create original text from a prompt. The free version of Grammarly does not use generative AI. Instead, it uses patterns it has learned from "reading" huge datasets of proper grammar use.
Representatives from Grammarly told a Fox affiliate that suggestions for corrections "are not powered by generative AI and [its suggestions] can still be accessed even when generative AI features are deactivated." However, Grammarly does have a premium version that uses generative AI, similar to ChatGPT-4.
For lawyers, then, the first step is to determine if the version of Grammarly you are using is powered by generative AI. If you are using generative AI, you may need to disclose it if you used it to create text in any document submitted to the court, depending on the court's rules.
Grammarly's Privacy Policy
Generative AI typically "learns" by reading vast amounts of data, including user input. The California Lawyers Association has warned that generative AI may share queries with third parties and advises that lawyers “must not input any confidential information of the client into any generative AI solution that lacks adequate confidentiality and security protections.”
Does this apply to Grammarly? According to its current privacy policy, Grammarly does collect user information, including "all the text and documents that you upload, enter, or otherwise transmit when you use our products."
However, it is not a keylogger, meaning it does not track all activity on your device, and it appears from its privacy policy that you can manage and delete your data to some extent. You can see a list of third parties Grammarly shares data on its website.
Lawyers in jurisdictions that have guidance on client confidentiality for AI should therefore be careful that any spell-checker or other AI algorithm has enough security protocols in place to avoid any client confidentiality concerns.
Related Resources
- Pro Se Litigant Fined 10k for Filing AI-Generated Reply Brief (FindLaw's Practice of Law)
- Rapper Pras Michel Contests Conviction Because of Lawyer’s Use of AI (FindLaw's Practice of Law)
- Bumpy Road Ahead for All in Adoption of AI in the Legal Industry (FindLaw's Practice of Law)
