I would hope any modern ai would be able to pass an exam as it’s regurgitation of information (especially with essays). Assuming proper training, it should be

I would hope any modern ai would be able to pass an exam as it’s regurgitation of information (especially with essays). Assuming proper training, it should be able to pass anything that’s within its training set. All the usual standards, such as “was hamlet mad?”, “was Hitler inevitable?”, “did Weimar Germany deliberately hyperinflate?”, etc. Things that have been boring ta’s since ancient times. It would be way more interesting to see what happened when moving beyond its training set - would it smoothly lose reasonableness or fall off a table. And to see how it’s lack of understanding word meaning trips it up in such extensions.