Grammarly has set the record straight about how it uses customer data for AI training and what is involved in opting out.

As we reported at WPN, Grammarly Support told Calibre’s Karolina Szczur that the only way for customers to opt out of their data being used for AI training was to purchase 500 licenses. For everyone else, Grammarly said it would “sample snippets of text at random, disassociate them from a user’s account, and then use a variety of different methods to strip the text of identifying information….In other words, we don’t store any text in a manner that can be associated with your account or used to identify you or anyone else.”

Following publication, Grammarly reached out to WPN to clarify its policies regarding AI training. According to a company spokesperson, “when businesses or organizations go through a security review with Grammarly, if requested, that business can opt out of Grammarly training on their de-identified, anonymized, and sampled data—opt-out is not limited to a 500+ license size.”

According to the company, there is no cost associated with requesting a security review, but only companies can do so at this time. Individuals licensees currently have no option to opt out of the AI training, although the spokesperson said Grammarly is planning to roll out more control for individual users.

More advanced user privacy and transparency controls for individuals are coming….All Grammarly users can view all data associated with their account at any time by requesting a personal data report.

It is unclear if the original support response Szczur received was a mistake or if the company has altered its stance in response. Either way, it’s good to see the opt-out barrier lowered.

At the same, as we have stated at WPN many times over the years, it’s one thing for a company to use customer data when providing a service for free. The customer’s data is an understandable trade-off for a free service.

It is an entirely different thing for a company to use a customer’s data, when said customer is paying for the service, and not providing a method to opt-out. The fact that the data is “de-identified, anonymized, and sampled data” sidesteps the fact that a customer should have control over their data when using a paid service.

Hopefully, Grammarly will deliver such an option to individual customers sooner rather than later.