"AI" is already handling the search for you. The big search engines are probably the first mass scale adopters of machine learning.

And they have lost the war with SEO spam to a hilarious extent. What makes you think the same won't happen with chat bot AIs? Bad actors (including PR agencies) will inevitably figure out where and how to spam comments in order to bias the AI models in favor of their agendas or products.

If the data they consume is filled with something like "fossil fuels don't cause global warming because XYZ", the chat bots will repeat it. They don't have the capacity to reason.

There hasn't been a reason to flood the internet with low effort spam because it's easily detected by humans who read it. But the ML algorithms will be a lot easier to trick.