I get that you are using algorithm as a short-hand here; but can you elaborate and describe what type of algorithms you suggest be opt-outable?
Algorithm is a super general term, but everyone uses it to describe something super-specific. I feel like everyone sorta has their own personal definition. I am in favor, at least in principal, of legislation that would curb some of the more abusive uses of datascience; but the idea of trying to legislate "algorithms" seems... fraught.
Agree it is hard but important to do to avoid discrimination. It’s less “regulating algorithms” and more “protecting personal data and access”.
Google and other big data co’s should not be allowed to profile unregulated and that’s exactly what they are doing.
I think a reasonable solution is that if a company is using someone’s “personal data” (legal definition pending) to configure/curate/personalize/uniquely-alter their experience in any way, that they allow the consumer or user to change those settings.
Configuration variables will open up transparency and is the first step to having a personal data right.
If an “Algorithm” is too complex to allow a user to configure (black box), they need to provide a way to opt-out to a general organic / main audience one then.
Algorithm is a super general term, but everyone uses it to describe something super-specific. I feel like everyone sorta has their own personal definition. I am in favor, at least in principal, of legislation that would curb some of the more abusive uses of datascience; but the idea of trying to legislate "algorithms" seems... fraught.