Let's try to move some discussions here.
I was just notified about this study: https://web.archive.org/web/20250509133900/https://www.science.org/doi/10.1126/science.adq1814
Basically it tested whether an AI chatbot can convince people to stop believing in conspiracy theories, and it did:
To evaluate this possibility, we leveraged developments in generative artificial intelligence and engaged 2190 conspiracy believers in personalized evidence-based dialogues with GPT-4 Turbo. The intervention reduced conspiracy belief by ~20%. The effect remained 2 months later, generalized across a wide range of conspiracy theories, and occurred even among participants with deeply entrenched beliefs.
Imagine receiving personalized propaganda on any big tech social media. Imagine this spilling over into schools, or maybe employee training. Imagine this eventually being integrated with video "influencers" (which will also be AI generated). Then imagine them being outside, on billboards, advertising "official versions" of everything. And later maybe merging with AR, so everyone will get their perfectly convincing personalized propaganda right into their eyes. Creating a popularion of perfect pawns that don't even realize they are such, because that is all they ever knew.
How about that for a conspiracy theory? But anyway, I really think this is a likely scenario if we don't stop it. The elites have no moral or emotional blocks as shown during Operation Corona.
I also wonder if the same effect could be realized but in the opposite direction. But, the LLMs would have to be trained on the conspiracy data then, and the "official" LLMs surely aren't.
Full study is downloadable here