53·
4 days agoIntellectual property is made up bullshit. You can’t “steal” a jpeg by making a copy of it, and the idea that creating something based on or inspired by something else is somehow “stealing” it is quite frankly preposterous.
The sooner we as a society disabuse ourselves of this brainworm the better.
Edit: I have very mixed feelings about so-called generative AI, so please do not take this as a blanket endorsement of the technology - but rather a challenge on the concept of “stealing intellectual property,” which I unequivocally do not believe in.
I will preface this with my usual disclaimer on such topics: I do not believe in intellectual property (that is, the likening of thought to physical possessions). I do not think remixing is a sin and I largely agree with the Electronic Frontier Foundation’s take that “AI training” may largely be fair use. So, I don’t think so-called “generative AI” is inherently evil, however in practice I think it is very often used for evil today.
The most obvious example is, of course, the threat to the work force. “AI” is pitched as a tool that can replace human workers and “wipe out entire categories of human jobs.” Ethical issues aside, “AI” as it exists today is not capable of doing what its evangelists sell it as. “AI chat bots” do not know, but they can give off a very convincing impression of knowledge.
“AI” is also used as a tool to pollute the web with worse-than-worthless garbage. At best it is meaningless and at worst it is actively harmful. I would actually say machine generated text is worse than imagery here, because it feels almost impossible to do a web search without running into some LLM generated blog spam.
Creators of “AI” systems use scraper bots to collect data for training. I do not necessarily believe this is evil per se, but again - these bots are not well behaved. They cause real problems for real human users, far beyond “stealing jpegs.” There is a sense of Silicon Valley entitlement here - we can do whatever we want and deal with the consequences later, or never.
I have long held that a tool, like any human creation, is imbued with the values and will of its creators, and thus must serve both the creator and the user as its masters (The software freedom movement is largely an attempt at reconciling these interests, by empowering users with the ability to change their tools to do their bidding). In the case of “Generative AI” it is very often the case that both the creators and users of these tools intend them for evil. We often make the mistake of attributing agency to these computer programs, so as to minimize the human element (perhaps, in order to create a “man vs machine” narrative). We speak of “AI” as if it just woke up one day, a la Skynet, in order to steal our jpegs and put us out of work and generate mountains of webslurry. Make no mistake, however - the problems with “AI” are human problems. Humans created these systems in order for other humans to use, in order to inflict harm to other humans. “AI slop” was created specifically for an environment in which human-generated slop already ran amok, because the web as it existed then (as it exists today) rewards the generation of slop.