I hope we both understand we already knew this.
My question is in two parts:
First: Philosophical concerning the entire concept of privacy, do you want a world in which your bar tender doesn't know your usual drink because you exercise your right to privacy by default.
(This example is a metaphor, please extract)
Second: AI specific, if we all adopt that approach, then your intelligence, knowledge and individual construct never gets added to the collective consciousness. Discuss.
On your first point: it’s good to have the ability to selectively disclose information. For example, if you’re inputting health data to get a diagnosis, you may want to keep it private. Or you’re writing a book and don’t want your IP used for the benefit of others. But yeah, there’s plenty of info you don’t mind sharing with the world.
Another consideration is that once you feed personal data to a public LLM, you can’t take it back. It literally gets incorporated into the model so there is no way of undoing it.
Agreed with your first point.
Your second point, also agreed, but isn't that true for life itself.
We all have things or aspects about us in the semi-public domain (perhaps just friends and family) that we'd prefer wasn't known.
As for collective data and privacy. The concept of Big Brother is both true and not true.
Yes, everything you don't want known about you is mostly known by Big Brother. However if Big Brother knows everything about everybody that they didn't want know, then whats the advantge.
For example the scam where you get an email saying "You've been captured watching porn, if you don't send X amount, then we will release that to the world".
Well generally, you have been watching porn, but the scammer doesn't know that, but everybody knows it probably true, so where's the scam?
Even if the scammer had the video of you watching the porn, so what?
The scam wins because you have been watching porn and society treats it as a taboo, which it isn't.
Remove convention and you remove the scam.
These are important but difficult questions, as technology bleeds into more domains, and governments/corporations are vying for all the data they can get their hands on.
I think it’s important to:
A. Have access to tools that have privacy as a default with the ability to selectively disclose information.
B. Provide transparency on how your data is being used. Example: when we go to a financial advisor or doctor, we take for granted that they’re not broadcasting our data for profit, otherwise they’d face legal action. But for the AI versions of these roles, there is a trade-off and people should be made aware of how their data is being used.
