this post was submitted on 11 Jun 2024
38 points (93.2% liked)

Tech

464 readers
1 users here now

A community for high quality news and discussion around technological advancements and changes

Things that fit:

Things that don't fit

Community Wiki

founded 9 months ago
MODERATORS
Vacant@programming.dev
38
Craig Hockenberry: "Apple’s model training data is either licensed or publicly available on the Internet" - Mastodon (mastodon.social)
submitted 5 months ago by sabreW4K3@lazysoci.al to c/tech@programming.dev
4 comments fedilink hide all child comments
 

I’m surprised this wasn’t mentioned in the keynote or State of the Union, but Apple’s model training data is either licensed or publicly available on the Internet. No personal information is used.

I don't think it's as ethically sound as the author thinks, but it's worth a discussion at least.

top 4 comments
sorted by: hot top controversial new old
[–] huginn@feddit.it 9 points 5 months ago (2 children)

They're partnered with OpenAI though, which is none of those things.

[–] sabreW4K3@lazysoci.al 5 points 5 months ago

This is why I said i don't think it's ethically sound. Just because I post something publicly, doesn't mean I'm consenting to it being used to train LLMs

[–] bokherif@lemmy.world 2 points 5 months ago

I'm not sure how extensive the partnership is, but the AI features (apple intelligence lol) are coming to devices with A17 pro (iPhone 15 pro lineup and later) and M chips. I believe the rest of the devices will use ChatGPT and they even mentioned that Siri will ask you each time if you want to use ChatGPT for the response. So ChatGPT partnership might only be for older devices.

[–] DarkThoughts@fedia.io 8 points 5 months ago

"Publicly available on the internet" is very wishy-washy. Just because something is publicized on the internet, does not mean you have the rights to use it. I hate how companies constantly get protected through draconian copyright laws but your average Joe is just content to be used to make even more money.