Precisely https://www.theguardian.com/technology/ng-interactive/2025/jun/27/tony-cox-epidemiology-risk-assessment-chatgpt-ai

conjure up an email summary within seconds that can shave off up to 5 whole minutes

… but can it? Like actually, can one do that?

Sure an LLM can generate something akin to a summary. It will look like it’s getting some of the points in a different form… but did it get the actual gist of it? Did it skip anything meaningful that once ignore will have far reaching consequences?

So yes sure an LLM can generate shorter text related to what was said during the meeting but if there is limited confidence in the accuracy and no responsibility, unlike somebody who would take notes and summarize potentially facing negative consequences, then wouldn’t the reliance on such a tool create more risk?

Interesting times, pardon the pun.

Actually… if you flip it…

So I’d rather argue that hammer can even screw screws.