There have been findings on political bias in LLMs, most prominently observed in US politics. We could try replicating these results to see what its left-right leaning is for non-Western democracies (Brazil, South Korea, etc.)
No answers yet
It would be interesting to compare the results with data like literacy rates. If for a long time only an elite had the means to write down their version of events, and that was used as training data, you could expect to see a bias for that elite in the LLM.
I think left-right leaning is the wrong axis for this. Instead it could be more interesting to focus on specific issues, debates, discussion of ambigously understood events, etc.