I suggest the summarize feature would be improved if it picked up data from the most frequently highlighted data. Imagine the summary included the top 3 quotes etc. It would put the reader more in touch with the primary text.
I would prefer that such a change be under the name condensation rather than summary. On a personal note, I would never use such a condensation because I've only highlighted (since college in the late 60's) items that are new to me or that I disagree with. Because of my own practice, I would have trouble believing that others' choice of highlighting was useful to me.
Note: I recently purchased a used copy of a Longacre textbook that was sold as having no markings. In fact, the first few pages were heavily marked in a manner that showed me that some poor student was in way over his head. I assume he dropped out of the course after a couple of days. His highlight would also be meaningless to me. Highlighting is sometimes surprising in what it tells you.
Maybe the AI would be smart enough to recognise truly valuable citations if you don't trust it to the crowd? In theology as in many cases its not just what is said but how it is said and if the purpose of the summary is to help us understand something quickly then helping quickly identify the most key passages is a useful tool to me. So many books contain so much waffle.
helping quickly identify the most key passages is a useful tool to me
I see your point but that is not how I would go about it. I'm not saying it isn't useful to many users. I'm saying it should be separate from the summarization as it destroys the confidence I have in summarization and meets the definition of condensation rather than summary as I understand the terms.
