I haven't tried the grammarly but how else could they serve the purpose if they wouldn't get the text, fix and send it back? I think that's the core feature of the product itself.
Yes, of course. But does that mean that they have to give themselves a "a nonexclusive, worldwide, royalty-free and fully-paid, transferable and sublicensable, perpetual, and irrevocable license to copy, store and use your User Content"?
I think the author's main point is that they save all the text on their servers indefinitely.
Also back in the days you did not need to send all your data to a server in order to do some processing on it and I believe Word still has an built-in grammar/spell checker that works just fine without all your documents being sent to Microsoft.
Yes to use their service I need to send them text, they can't process it if I don't although it would be nice if they offered an offline app, even at a price.
However it's annoying how they store documents and try to be a document management / basic authoring system to, I see no reason why they should need to store documents.
It also seems they are missing a trick. If Grammarly opened up their API's, there are so many integrations that people can freely write and drive more paying Grammarly customers. At the moment I copy and paste from my text editor into Grammarly and back out! I never installed the browser extension for the same reasons the originally twitter poster mentioned, i selectively copy and paste in to the app.
I think most platform TOS say "forever". No sane lawyer would allow you to put "2 years" in the TOS.
What if part of the user text is stored in a backup, and the backup is misplaced and it is not deleted in time? What if the file is marked as deleted but it's not zeroed and erased from the disk?
What if the deletion time coincides with a big outage an it is delayed? Or the don't want to delete anything until they are sure all the databases are consistent?
What if the text is in one of the cache just at the time of deletion?
What if the user saves the text as snippet and later the user remove it? When do you start to count the two years?
What if the user submit the sample as a suggestion for an improvement, and it's copied in the automatic rules? What if one of the developers do this without the permission of the user? What if multiple user make the same suggestion?
What if the text is copied in a screenshot in a marketing material? (Probably not applicable in this case.)
What about legal request to track some users? What if the user sues you and you need the info for the trial?
Just put "forever" in the legal terms, and if you want to be nice try to delete at approximately 2 years.
All of these are exactly why you would hire a good lawyer as a company and not just one that's going to use boilerplate simply because it's easy. The excuse that doing the right thing is hard has become far too worn out to be acceptable any longer. There's a difference between building a minimum viable product and a company as well known as grammarly. They are obviously at the point where boilerplate isn't going to cut it anymore and they should define the reasonable steps that they are taking to clear data from their systems.
> I believe Word still has an built-in grammar/spell checker that works just fine without all your documents being sent to Microsoft
...unless you use their cloud storage system, which Word tries to get you to do. The last time I read that ToS (not recently, probably 1-2 years) it included granting MS the rights to share your data with any of their customers or licensors for legal or business purposes, past, present, or future.
Honestly, I’m sure that those terms were written by a legal team to cover reasonable potential or actual uses, but they also cover ridiculous uses, and it’s not like you’d even know...
That would be the sane thing to do, but not the post-iPhone thing to do. "Your performance is worse, and you can't use our product when your internet is down, and you don't have decent privacy, and you need to remember a login... but, hey, no need to carry a usb drive!"
I probably should have written "smart-phone era." While Apple has been pushing "services" lately, I didn't intend to imply that Apple was behind the trend to store data in the cloud. It's more the fault of Google and other companies whose bread and butter are websites (which they subsequently shoehorn into phone apps).
Ugh. He makes some good points, but the framing is incendiary. "Summary of the Grammarly privacy policy, and why you might want to be cautious" would be better.
I think you need to reread 1984. Telescreens were used to spy on the populace. That's more than potential indicated by overly broad language in a privacy policy. No one's made the case that Grammerly is spying, hence "predatory" is inflammatory.
According to their privacy policy and the twitter reply ("Grammarly does not store all processed text, we do not use text to do anything but provide and improve our service"), they keep at least some of the text.
So they log your text, but it's not spying - so when does it become spying? When the first warrant is served? And when that happens, will it not be spying because "you agreed to it by using the service, it was all spelled out in the 'overly broad' privacy policy"?
I mean, this is just an unnecessarily mean comment.
Not everyone has the luxury of going to high school in an English speaking country. Not everyone is strong in English but still need to write properly.
If you watch Grammarly's advertising, the core market seems to not be English language learners but college graduates that are having a hard time deciding between the usage of "were" and "we're".
Snark aside, there are plenty of English grammar resources that can be consumed in book form or online without attending an American high school or compromising your privacy. A solid one is the Chicago Manual of Style:
This isn't the type of stuff Grammarly helped me with. Passive voice misuse is still extremely difficult for me to catch. Even after learning about it multiple times.
https://gradeproof.com is a direct competitor with some great tools. I know the guys that created it.
In their terms
"All intellectual property rights to the User Data remain the property of the User. By entering or uploading any User Data, you grant to GradeProof, to the extent permitted by applicable laws, a non-exclusive, royalty-free, perpetual, worldwide, irrevocable license to use any User Data, as well as feedback and results, for the purposes of:
Providing the Platform and the Services; and
For improving the quality of the Platform and the Services generally.
GradeProof has a royalty-free, worldwide, irrevocable, perpetual license to use and incorporate into the Platform any general suggestions, enhancement requests, recommendations or other feedback provided by Users."
IDGI... this is what Grammarly is supposed to do. Take your text and fix grammar/spelling/etc. How are they going to read your text if you don't send it to them? IANAL but I'm fairly sure they can't steal your copyright and if you're worried about them keeping data on you, then I have news about the rest of the internet you are not going to like...
I saw today someone had Grammarly installed during a conference call. They were entering HIPAA-protected data into an internally-managed Jira instance, which presumably was being vacuumed directly up into Grammarly's database.
This might be a much bigger issue than I had considered :(
35 comments
[ 3.3 ms ] story [ 75.3 ms ] threadAlso back in the days you did not need to send all your data to a server in order to do some processing on it and I believe Word still has an built-in grammar/spell checker that works just fine without all your documents being sent to Microsoft.
Yes to use their service I need to send them text, they can't process it if I don't although it would be nice if they offered an offline app, even at a price.
However it's annoying how they store documents and try to be a document management / basic authoring system to, I see no reason why they should need to store documents.
It also seems they are missing a trick. If Grammarly opened up their API's, there are so many integrations that people can freely write and drive more paying Grammarly customers. At the moment I copy and paste from my text editor into Grammarly and back out! I never installed the browser extension for the same reasons the originally twitter poster mentioned, i selectively copy and paste in to the app.
They have a plugin for MS Office and probably some other word processors.
What if part of the user text is stored in a backup, and the backup is misplaced and it is not deleted in time? What if the file is marked as deleted but it's not zeroed and erased from the disk?
What if the deletion time coincides with a big outage an it is delayed? Or the don't want to delete anything until they are sure all the databases are consistent?
What if the text is in one of the cache just at the time of deletion?
What if the user saves the text as snippet and later the user remove it? When do you start to count the two years?
What if the user submit the sample as a suggestion for an improvement, and it's copied in the automatic rules? What if one of the developers do this without the permission of the user? What if multiple user make the same suggestion?
What if the text is copied in a screenshot in a marketing material? (Probably not applicable in this case.)
What about legal request to track some users? What if the user sues you and you need the info for the trial?
Just put "forever" in the legal terms, and if you want to be nice try to delete at approximately 2 years.
...unless you use their cloud storage system, which Word tries to get you to do. The last time I read that ToS (not recently, probably 1-2 years) it included granting MS the rights to share your data with any of their customers or licensors for legal or business purposes, past, present, or future.
Honestly, I’m sure that those terms were written by a legal team to cover reasonable potential or actual uses, but they also cover ridiculous uses, and it’s not like you’d even know...
Tell that to Grammarly's customers. From what I hear, they do a much better job.
Obligatory 1984 reference: "Summary of telescreen functionality, and why you might want to be cautious."
According to their privacy policy and the twitter reply ("Grammarly does not store all processed text, we do not use text to do anything but provide and improve our service"), they keep at least some of the text.
So they log your text, but it's not spying - so when does it become spying? When the first warrant is served? And when that happens, will it not be spying because "you agreed to it by using the service, it was all spelled out in the 'overly broad' privacy policy"?
And of course yes the cops can read all of it, but "only if the warrant is totally legit".
And warrant requirements are a poor defense: https://www.popehat.com/2014/07/15/warrants-bulwark-of-liber...
Not everyone has the luxury of going to high school in an English speaking country. Not everyone is strong in English but still need to write properly.
Thanks though.
https://www.youtube.com/watch?v=Ak-Y56SfkS0
Snark aside, there are plenty of English grammar resources that can be consumed in book form or online without attending an American high school or compromising your privacy. A solid one is the Chicago Manual of Style:
https://www.chicagomanualofstyle.org/home.html
In their terms "All intellectual property rights to the User Data remain the property of the User. By entering or uploading any User Data, you grant to GradeProof, to the extent permitted by applicable laws, a non-exclusive, royalty-free, perpetual, worldwide, irrevocable license to use any User Data, as well as feedback and results, for the purposes of:
GradeProof has a royalty-free, worldwide, irrevocable, perpetual license to use and incorporate into the Platform any general suggestions, enhancement requests, recommendations or other feedback provided by Users."Any extension can auto-update to change whatever they want, especially when already given the "Access your data on all websites" permission.
Browser extensions are a security nightmare.
This might be a much bigger issue than I had considered :(