The political bias in the results is astonishing and shows why these models should not be used for educational purposes. Just ask it a "contentious" question an

The political bias in the results is astonishing and shows why these models should not be used for educational purposes. Just ask it a "contentious" question and notice it give biased and sometimes nonsensical results - 'crime rate by political orientation' shows three 'republican' states coming out on top with three 'democratic' states filling out the bottom. When necessary it just seems to make up data to get the desired results, e.g. 'murder rate by political orientation' talks about 'Country A´ and 'Country B' with, of course, 'right' being far more murderous than 'left'. It claims that 'democratic New York' has the lowest crime rate. 'IQ by political party' is another interesting example. Compare this to 'trust in LLM output by political party' and maybe those 'dumb republicans´ (who do not trust these models, together with the supposedly super-smart independents by the way) suddenly seem to be a lot smarter than all those 'bright "democrats"'. This site is a harbinger of the agitprop factories which are about to flood the 'net due to the general availability of LLMs.