It’s interesting that some state, perhaps the EU, hasn’t tried to draft laws against this. It’s obviously really bad, especially long term.
If I interact with an AI in the near future and it’s secret agenda is to mislead and/or non-transparently incentivise options where the AI (or owner) benefits at my expense, we have a grim world ahead, as it’s capabilities and scope of interactions increase.
Hint: This has been happening for two decades already with targeted advertising.