Its amazing that such a sober take on the issue is getting downvoted without a single counterargument
Probably because the comment can be mostly summed up to: 'The majority of humans are idiots who need to be protected from themselves by the government.'
I am in fact saying “the majority of humans are not capable of defending themselves against every possible criminal, and therefore that responsibility belongs to the government.”
> When Peter put his contact information online, it had an intended context of use. Unfortunately, applications built on top of GPT-2 are unaware of this context, and might thus unintentionally share Peter’s data in ways he did not intend. That's expected when you publish anything online, you lose control over the data.
So copying an internet-published trailer of a Marvel movie is a crime, but my privacy is up for grabs.
Depends entirely on what you share publicly. GPT's training data involves only publicly crawlable dataset.
We should be more concerned about mitigating GPT-like systems because one thing is for sure, we can't stop these models. The genie is out of the bottle and I'm sure multiple actors are working on similar systems.
Isn't that how humans work? Sometimes I wonder if I have free will and sentience, or if I am just stringing words together based on some probability function. I've no doubt taken into account copyrighted works and personal information when training my built-in neural network. The examples the article gives like "misremembering" the murder as the murder victim sounds like something a person would do. Knowing verbatim…
It’s my understanding that under our current scientific understanding of the mind, we cannot have “free will”. Where would it come from? All science knows is that neurons receive inputs from sensors, mixes them together (with some quantum randomness sprinkled in), and eventually actuates muscles to produce outputs, exactly as GPT. Thus “free will” must come from some sort of faith. Either it was given to us by “God”,…
Free will is a cultural concept, not a biological one. Consciousness may be driven by a biological machine, but development of our ego is influenced a lot by culture.
Free will is faith in the sense that we learn the concept from our culture as our sense of self develops. The assumption of having it shapes our thinking such that we can say we actually have it.
A person who took their inner voice as voice from the gods would have a restricted free will in this view.
I was expecting a bloom filter based solution to block verbatim reproduction of training data. They just need to hash the sensitive n-grams (hopefully, a small part of the whole dataset) and store one bit per hash. Alternatively, they could do something like GAN and have a 'discriminator' classify if a sample is natural or synthetic. Then, at inference, condition to be original. So, verbatim training data reproductio…
If you want a model that returns original results, you're looking for a database.
Probably because the comment can be mostly summed up to: 'The majority of humans are idiots who need to be protected from themselves by the government.'
It’s not that humans need to be protected from themselves (individually), but that they need to be protected from other unscrupulous people , because computers are empowering the unscrupulous (and even the accidentally unscrupulous) more than they are empowering the common person. I broadly agree with ojnabieoot.
So why not teach people so they can empower themselves instead of making them rely on yet other people, who may or not be scrupulous themselves for protection?
Because the government is always scrupulous and trustworthy...and nobody that's ever said trust us we'll protect you has lied...
> There is a legal grey area as to how these regulations should apply to machine learning models. For example, can users ask to have their data removed from a model’s training data? Moreover, if such a request were granted, must the model be retrained from scratch? The fact that models can memorize and misuse an individual’s personal information certainly makes the case for data deletion and retraining more compelling.
This is an interesting angle I had not considered before. It seems like “right to be forgotten” requests could be quite damaging to the “train once run anywhere” promise of some of these models. (Or this could just mean that the training data needs to be more carefully vetted for personal data, but probably both as no vetting process can be 100% successful).
It’s not that humans need to be protected from themselves (individually), but that they need to be protected from other unscrupulous people , because computers are empowering the unscrupulous (and even the accidentally unscrupulous) more than they are empowering the common person. I broadly agree with ojnabieoot.
So why not teach people so they can empower themselves instead of making them rely on yet other people, who may or not be scrupulous themselves for protection? Because the government is always scrupulous and trustworthy...and nobody that's ever said trust us we'll protect you has lied...
You can try that, but that approach tends to fail, because people don’t like changing what they do, and they definitely dislike jumping through extra hoops. You can improve things with education, but only up to a point. Most people just don’t want to think about any of these things, so until it gets really bad (most commonly of the “leopards ate my face” variety, when something you’ve tacitly endorsed suddenly starts to harm you), they’ll normally just go along with whatever happens. This is the whole reason for government, to maintain rough balance so that no one gets too upset. It’s extremely imperfect, as you correctly imply, but that’s the general idea of it, and it normally works well enough (better than nothing, at least), and no one has presented any better alternative.
Maybe this data will be used to 'upload' you in a virtual world after the singularity. It might be your second life.
When thinking about this, you have to ask, who will profit, and how? It doesn’t seem like a clone of your brain running in a simulation is going to be of much advantage to you from the current industry incentives. Maybe if you’re rich?
AGI might be running our models just like you run your docker containers. Maybe it investigates other ways humans could have invented it by recreating the period before the singularity.
If you make some fairly plausible assumptions about Moores law (and equivalents), the simulations are cheaper than the meat bodies before mid-century. • https://worldbuilding.stackexchange.com/questions/51746/when... • https://kitsunesoftware.wordpress.com/2018/10/01/pocket-brai...
That is cool, but tangential to the question posed above - given that every part of surveiliance capitalism is currently working against my interest, either to sell me something I dont need, to inflienxe the way I vote, etc. Will the simulation of my brain belong to a megacorp, and be used to do even more damage?
If you make some fairly plausible assumptions about Moores law (and equivalents), the simulations are cheaper than the meat bodies before mid-century. • https://worldbuilding.stackexchange.com/questions/51746/when... • https://kitsunesoftware.wordpress.com/2018/10/01/pocket-brai...
That is cool, but tangential to the question posed above - given that every part of surveiliance capitalism is currently working against my interest, either to sell me something I dont need, to inflienxe the way I vote, etc. Will the simulation of my brain belong to a megacorp, and be used to do even more damage?
There is simply no way corporate hegemony will allow you to own the rights to the digital copy of you that they built using the tools of surveillance capitalism. No corporation or lawmaker will grant you rights over corporate property, even if that property was obtained through questionable means and/or your consent was never actually received. That's why these data snooping agreements corporations force into their products reserve the right to do whatever they want with your data.
The obvious response to this is, "Then don't patron those companies," which is just as naive and oversimplified of a solution as the one above suggesting everyone everywhere should just use encryption (but for very different reasons).
There is just simply no logical reason to believe consumers would ever get data rights over corporations under capitalism with major reforms putting corporations back under our boots where they belong.