181–190 of 514 posts
Earlier quoted context omitted.
Not telling you whats wrong should be highly ilegal, given how important these systems can be for modern society.
Isn't it already under GDPR where entity using AI to make decisions need to provide an explanation how it arrived at it?
I'm not defending Meta and I agree that AI shouldn't used to review appeals, but the title is clearly a clickbait. Based on the article, they banned you because AI thought you're trading live animals not because you're teaching Python.
So I'm curious to see the original title and ad image.
The whole issue around not telling you what rule you broke "because then you might work around it" is total BS. It's like a company I complained to the other day because they didn't label one of the inputs on their password reset form because "it would help a hacker know what to type in it" (it was the password reset password they expected but you wouldn't know). How much quicker would this be resolved if the first A…
The review loop should at least have gone to a person. It’s hard to assume both review loops missed something like that.
I have never put ads on FB - I would assume that there is at least a tagging, or topic orientation for ads (Education, Coding, CS, IT, for example).
The fact that the full ban came so quickly after the appeal is also very strange.
Earlier quoted context omitted.
To be somewhat fair to Meta; the "violation of unclear rules part" part is an appropriate response for spam detection . Usually any entity dealing with the web don't want to reveal how their spam detector works because doing so could in turn reveal information on how to circumvent it[0]. The issue is that Meta thinks they're large enough that they can apply this to every single one of their business customers and all…
Or they could, you know, spend some of their billions on employing humans to review these cases.
Which is why it seems really strange that this was a lifetime ban, and so quickly at that.
Maybe it didnt even get to human review?
At no point does the author provide a single example of an advert they created. Does no-one else find this slightly fishy?
Some were advertising my paid courses on Python and Pandas.
Others were advertising my free courses about regular expressions.
Still others were advertising my mailing list.
None of this comes even close to anything illegal, unless I'm massively mistaken.
Earlier quoted context omitted.
And some of these companies pride themselves as AI technology pioneers. Yet it seems their automated systems in this case are simply doing keyword matching. Mind-boggling!
Another sad anecdote: A bank in Finland suspended an account for a text in free-form message field of a bank transfer. The message was for a vet visit of a dog named Ira. When you say "Ira's payment" in Finnish, you add suffix -n, so the text field said Iran, which of course must indicate money transfer to a sanctioned country. It is comforting to know that the bank system catches all the illegal activity that the sa…
Being a bit stubborn, he still kept the message and just applied rot13 to it.
Earlier quoted context omitted.
And some of these companies pride themselves as AI technology pioneers. Yet it seems their automated systems in this case are simply doing keyword matching. Mind-boggling!
Honestly I’m not so sure it would help, as these bots would be trained on previous customer service interaction. I’ve been trying to get an Instagram account deleted for the better part of a year, and even with quintuple (!) ID verification and threatening to bring an GDPR case to bear, all I get is what amounts to “computer says no”, in such a talking-in-circles kind of way I am not sure I am talking to a human.
In case that leads nowhere, you should raise a complaint with you local data protection agency, for example https://ico.org.uk/make-a-complaint/