Yes. We all train chatbots the same way. There's a first step where it's fed a lot of text to learn to mimic coherent language and learn facts. And then there's a second step where we tune it to do the thing we want it to do. For example write dialogue, answer questions... In contrast to for example write a novel or write unproductive Reddit arguments.

All of this happens disregarding of who does it. A commercial business or a non-profit. And you'd be left with a similar result with all the upsides and downsides of a chatbot. It will hallucinate, make people spiral into stuff, give bad advice... Amplify mental disorders. That's just what chatbots do, due to how they're made. And there's currently no way around it.

So it really doesn't matter all that much who does it. It'd help if we had a different type of chatbot which isn't subject to the same kinds of limitations. But we don't.