this post was submitted on 08 Aug 2023
658 points (96.1% liked)

Technology

55760 readers
2810 users here now

This is a most excellent place for technology news and articles.

Our Rules

  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots

founded 1 year ago
MODERATORS
L3s
[email protected]
[email protected]
[email protected]
enu
[email protected]
L4s
658
The only way to avoid Grammarly using your data for AI is to pay for 500 accounts (yiffit.net)
submitted 11 months ago by [email protected] to c/technology
37 comments fedilink hide all child comments
 

Source: https://front-end.social/@fox/110846484782705013

Text in the screenshot from Grammarly says:

We develop data sets to train our algorithms so that we can improve the services we provide to customers like you. We have devoted significant time and resources to developing methods to ensure that these data sets are anonymized and de-identified.

To develop these data sets, we 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 (such as identifiers, contact details, addresses, etc.). Only then do we use the snippets to train our algorithms-and the original text is deleted. 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.

We currently offer a feature that permits customers to opt out of this use for Grammarly Business teams of 500 users or more. Please let me know if you might be interested in a license of this size, and I'II forward your request to the corresponding team.

all 39 comments
sorted by: hot top controversial new old
[–] drdabbles 74 points 11 months ago (3 children)

There is a second way, legend has it. The ancient ones tell a tale of the one that does not use the service, and does not train someone else's shitty models for free.

[–] [email protected] 37 points 11 months ago (2 children)

I hate to break it to you, but we're all presently training someone else's shitty models for free by commenting on Lemmy. Probably multiple organizations at some point, in fact.

[–] drdabbles 23 points 11 months ago (1 children)

You didn't break it to me.

[–] essteeyou 25 points 11 months ago (1 children)

Yeah, when I write something public I accept that anyone can read it or use it for whatever reason. When I pay for a service then it's a bit of a grey area. When the service is free I know my data will be used to make money by any means necessary.

[–] [email protected] 7 points 11 months ago (1 children)

Thats why I make misatkes in everyting I write. They won't have a good set too take from my comments.

We should alll (w)right wrongs like this.

[–] [email protected] 2 points 11 months ago

Rigging wrongs by writing wrong, loivcjfrrrjdd it!

[–] Gimly 3 points 11 months ago (1 children)

You raise a very good point, people cannot be mad that companies use data that they made public to train their ai. It's public, people can do whatever they want with it. We really need to teach people to be more careful with what they post online.

But I'm wondering, is there a default license for data posted on lemmy?

[–] [email protected] 3 points 11 months ago

That would imply ownership and agency over the retention of our data, which federation kind of fundamentally cannot guarantee. An instance in the Fediverse can only guarantee the right to be forgotten on their own instance. I could see this becoming a big regulatory problem as the Fediverse grows. We're already seeing regulatory issues with CSAM, for example.

[–] [email protected] 10 points 11 months ago (1 children)

The third way is like the second way: we learn to write good without crutches.

[–] drdabbles 3 points 11 months ago (1 children)

As a native American English speaker, shit's mad hard yo.

[–] [email protected] 0 points 11 months ago

Native American English you say? You write-um heap good talk!

[–] FlyingSquid 3 points 11 months ago

Please read turn to page 32 of your copy of Strunk & White and recite the prayer against avoiding excess verbiage.

[–] [email protected] 44 points 11 months ago (1 children)

For anyone looking for an alternative, LanguageTool though not perfect, has shown itself to be far more privacy respecting

[–] [email protected] 7 points 11 months ago* (last edited 11 months ago)

Came here to say this. I've been using LanguageTool for a while, but they've also recently started implementing AI into the product.

Using AI itself isnt a problem if the engine they're using is completely proprietary. But they're likely using some third-party engine to send the data to. But I'd love to be proven wrong by them open sourcing the code for it so I can take a look at it myself.

[–] Nioxic 43 points 11 months ago

Well isnt grammerly not just a keylogger with some helpful features to put your mind at ease?

[–] art 31 points 11 months ago

The only way to avoid Grammarly using your data for AI is to pay for 500 accounts

Protip: You can also simply not use grammarly.

[–] brygphilomena 28 points 11 months ago (4 children)

Let's ignore the ethical implications of this for a moment.

Grammarly is training it's AI off of the poorly written grammar of it's users that it has to already adjust?

It seems like this would be a flawed set of training data. It's training on what it already either produced or on something written by someone who may not have used proper grammar in the first place.

Am I to expect this AI will improve over time?

[–] [email protected] 15 points 11 months ago

Depending on how they're training it, they're likely looking at when grammarly corrections were accepted or rejected and the context around that. That's what I'd be using from the dataset anyhow

[–] [email protected] 6 points 11 months ago

Remember that people have said GPT4 is getting dumber because of interacting with humans.

[–] [email protected] 2 points 11 months ago

On average, people's grammar is correct, kind of by definition.

[–] funktion 1 points 11 months ago* (last edited 11 months ago)

I can actually tell you a little about how this worked. The training data went through a team of grammarly writing experts before being fed to the AI. Writing experts would get short snippets out of context, often from publicly available text (e.g., reddit comments, classic novels, poems, scientific papers) and correct it for both grammar and clarity, then that would then become training data. Later on the team would do the same for content generated by the AI.

Source: Was one of the writing experts. For a couple of weeks I was correcting snippets that were very very obviously from r/squaredcircle. Very weird reading about Dave Batista's giant dong at work.

[–] [email protected] 26 points 11 months ago

Grammarly is basically a keylogger anyway, with every stroke send to their servers. Why ANY business even allows their employees to use this is really beyond me.

[–] mysoulishome 22 points 11 months ago (3 children)

Why is it necessary to post this to so many different cimmunities?

[–] [email protected] 33 points 11 months ago (1 children)

I see this complaint a lot but honestly I don’t quite understand what the big deal is. Not everyone is subscribed to the same communities. Personally, I’d love a feature on kbin/lemmy that rolled up duplicate posts on the client, but it’s really not that annoying for me to see a couple dupes in my feed if they’re posted in relevant communities /shrug

[–] [email protected] 44 points 11 months ago

Case in point, this post is the only one I’ve seen of this, so I must not be following any of the other communities/magazines the OP posted to.

[–] [email protected] 14 points 11 months ago

Maybe it's relevant to those community i guess, so they posted it there. I didn't see such complaint in reddit where multiple sub posted the same stuff over a few days(eg: games, gaming, pcgaming, pcmr, so on and so forth, can share the same exact news and it will appear in your feed multiple times), so not sure why it suddenly a problem in Lemmy.

[–] Mentrix 20 points 11 months ago

I work in the field of data science and I really get why data is needed and why it makes sense to collect use data but enforcing this and now allowing to opt out free of charge is simply not okay

[–] Iron_Lynx 6 points 11 months ago

I'm quite certain this policy is illegal in some jurisdictions (read: the European Union).

[–] [email protected] 2 points 11 months ago* (last edited 11 months ago) (1 children)

So... grammarly is problem to help you write an email or a document you will send via gmail or publish online?

Or you use grammarly for private diary?

[–] gps 2 points 11 months ago (1 children)

Most people consider email private. Plus a lot of people use Grammarly for work documents.

[–] Gimly 2 points 11 months ago

The same people will use Gmail without batting an eye and we all know what Google has been doing for years with emails in Gmail.

It's quite funny that now that there are "AI" that everyone can use, people get all worried about their data being used to train them but nobody cared before when it was Google or amazon using them to train their models.

[–] uzziah0 1 points 11 months ago

So, we should create really bad grammatical texts and submit to grammerly, to befuddle it?

[–] Kinglink -1 points 11 months ago* (last edited 11 months ago) (2 children)

I mean you've been using their service for free...

Listen I get people who but they have said they train their programs based on your responses for a while. If your not paying for these services then yeah they are going to monetize and improve the service off of what you do and includes ai.

Start paying for the services you use... And I don't mean simple ten or twelve dollar a month plans, which I'm betting she doesn't use either

People really have gotten used to the free lunch model...

[–] [email protected] 19 points 11 months ago

they have a personal use premium service, which the author already pays for.

[–] puppy 4 points 11 months ago (1 children)

What do you recommend she do? Spend for 500 accounts?

[–] Kinglink 1 points 11 months ago (1 children)

Find better solutions. I found that grammarly results are similar to office. If you want to buy a product buying word versus renting grammarly would be a better choice.

There's multiple other websites that do the same as grammarly. But if you choose to go with a free option or low cost option this happens

[–] asamson23 1 points 11 months ago

As a Québécois, one piece of software I really like is called Antidote, which does pretty much the same as Grammarly, without the online components, unless you pay for the subscription. It also works in both French and English.