josefright now, when people talk about AI, we tend to really like to lump it together into this idea of "the AI", like "the algorithm" or whatever. we all know we're being facetious when we do that, it's just fun to say, an appealing meme. obviously there are a lot of different machine learning models and approaches, but mostly "AI" is used to denote this singular type of smart-yet-stupid limerick-writer and art plagiarist, presumably made in the image of its creator, the witless and amoral tech guy
i think over the next 10 years, we'll start to see a split in this, as most people become comfortable with all the different permutations of what we currently just call "AI", and become interested in comparing their motivations, capabilities, and cultural associations, and we begin to get a few different labels to describe the different types or vibes of AI systems. probably these will emerge from moral categories, and there'll be an AI equivalent of "greenwashing" to make your AI look ethical
i think there could be two axes used to compare AI systems: face/heel and open/proprietary. this nets us 4 potential types of AI, but they haven't been given names yet. the face companies will appeal to non-tech-aligned media orgs with their ethical selling points, the heel ones won't care about such limitations and will be more powerful as a result. the open systems will be less advanced than the proprietary systems but easier to deploy and extend, and might become successful purely due to that
face-proprietary AI will be used by tech startups, heel-proprietary AI will be used by military robots, face-open AI will be used by hobbyists, and heel-open AI will probably just be used to generate porn or harass people or whatever. of course we won't actually use these 4 names, we'll come up with something better and more 2020s
neo@jk virgin for proprietary and chad for open. then lawful/chaotic
@jk Well, the way I see it, the other axis has two extremes: Either the user comes first or society comes first. EG chaotic or lawful