Meta won't train AI on Euro posts after all, as watchdogs put their paws down
Meta won't train AI on Euro posts after all, as watchdogs put their paws down
Facebook parent calls step forward for privacy a 'step backwards'
common eu w
So Meta's AIs will mainly reflect non-EU cultural values.
EU cultural values include resisting against corporations doing whatever they want with our data. Let's see meta try to reflect those.
Mate, it takes the combined discussions of the internet
I would not expect any sort of values
I would expect plenty of deeply-held values:
Rule 34
Shitposts
CSAM
Disingenuous partisan mis/disinformation
Worst hot takes imaginable
Its crazy how much further ahead Europe is in Privacy Protection.
All these companies need to be held responsible for what they do with our data, and what it costs them when they lose control of it. Either figure out how to safe guard it or suffer painful consequences. Or perhaps only store what's necessary for us to interact.
But then again, we also have pretty much every EU group pushing for super invasive chat control. It's ridiculous how schizophrenic they are on the subject of digital privacy.
They are telling what they care about, take notice.
I am once they get a local AI grfiter, they will change tune too.
It's not the same groups and entities pushing these things. It looks contradictory because it all ends up submitted to the same legislative bodies but that's par for the course in a functional democracy.
Yeah, seems weird, but there's also points where it's not related at all.
One is a company using user data they didn't tell they would use for this purpose, and illegally trying to do it anyway. They literally sell the data by making a product of it. It's also a private company with stakeholders.
Other is EU scanning messages, but not selling them.
So it's about who you trust basically.
Don't worry, maybe Meta can eventually just buy the inevitable leaks resulting from the general chat surveilance the EU so vehemently tries to push through.
This is the best summary I could come up with:
And while this climb down has been cheered by privacy advocates, Meta called it "a step backwards for European innovation" that will cause "further delays bringing the benefits of AI to people in Europe."
"We're disappointed by the request from the Irish Data Protection Commission (DPC), our lead regulator, on behalf of the European DPAs, to delay training our large language models (LLMs) using public content shared by adults on Facebook and Instagram — particularly since we incorporated regulatory feedback and the European DPAs have been informed since March," the social network said in a statement on Friday.
Without a steady diet of EU information, Meta's AI systems won't be able to "accurately understand important regional languages, cultures or trending topics on social media," the American goliath said at the time.
"In order to get the most out of generative AI and the opportunities it brings, it is crucial that the public can trust that their privacy rights will be respected from the outset," Almond continued.
"We will continue to monitor major developers of generative AI, including Meta, to review the safeguards they have put in place and ensure the information rights of UK users are protected."
Privacy group noyb had filed complaints with various European DPAs about Meta's LLM training plans, and its chair Max Schrems on Friday said while the organization welcomed the news, it "will monitor this closely."
The original article contains 589 words, the summary contains 231 words. Saved 61%. I'm a bot and I'm open source!