While it's true that the first x characters of an email don't always have the most important bits, they're never _wrong_. It's always exactly what's in the email.

I'm worried about cases like this [0], where a summary is inaccurate or misleading. It may not happen often, but it's possible, so I have to be a little skeptical of every summary. Though I can see where AI summaries are a plus, the move from 100% accurate to <100% feels like a big downgrade.

[0]: https://mastodon.macstories.net/@johnvoorhees/11288700482217...