I suspect that would backfire immediately, as it would mean all stories are indirectly (via many people using this) filtered by what Gemini finds plausible.

Purely as an example, when Trump was making the threats against Greenland, I tried asking ChatGPT what the economic implications would be and how to hedge against them, and it had a great deal of difficulty with even the rhetoric being "current events" rather than "a fictional scenario that will never happen".