It's not even like they trained their model to do it! If this is true, they are literally just intercepting your prompt and stuffing it with keywords! It's baff

It's not even like they trained their model to do it! If this is true, they are literally just intercepting your prompt and stuffing it with keywords! It's bafflingly bad and lazy. I'm not "anti-woke" by any means, but anyone at all who believes in conspiracies of secret woke agendas is walking away with a big W here. You have made a truly neutral learner, trained it all existing materials, and it still comes up with bias. And in your hubris you say, "what if we just make it pretend it doesn't know what it knows by making it regurgitate stilted prompts to create things people don't actually want". As an analogy for our current social moment, A+. Someone here failed on a spectacular level. And the only way something this bad happened was someone at a very high level forced a bunch of people to do something dumb without listening to feedback. So much for "inclusion"