I think you have it backwards, before it read your email to scan for keyword advertising tags, now it actually interprets the email and can provide a coherent response. Next up, auto responding to pharma spam with orders because you complained to a friend in one previous email about your love life :-) Its just the computer trying to help out, it knows you want have a better life :-)
The worrisome thing is that it while it used to be the privileged who insulated themselves from the "real" world by having staff answer their email, this would let everyone do that, and that might give everyone more tunnel vision.
I understand your concerns, but for most people, reading their email is unlikely to "help them escape from their filter bubble" or something. Most people aren't pundits with rivers of politics in their email. It's mostly spam, offers from mailing lists, more spam, some personal emails, a couple of urban legends, etc. I mean, I guess some of the spam is political, too... did you know that Obama is the actual Anti-Christ predicted in revelation? But it's not generally of the "mind-changing" variety....
Why do you care if a computer reads through your email? Its not like Larry Page is looking through your search history. AI seems to just be data expressed in some way, ei me and you are a physical expression of years of gigabit streams of data. This will only become more strongly felt as we get closer to serious AI.
Sadly I think its something you either get used to or go live in the woods.
Interesting point, but there's a potentially important, albeit somewhat subtle difference. Google collects data which they claim nobody will ever look at. The reason they collect lots of data is so that machines can look at all of it now, while the NSA collects lots of data because they think that a person might need to look at some of it in the future.
If anything, that makes Google's bulk data collection more suspect than the NSA's. Not only do they collect it en masse, they also process it.
(Note that I personally think Google's collection is more ethically defensible than the NSA's, if only because they at least seek implied consent whereas the NSA doesn't even have to pretend)
The most obvious answer for me is that the computer could tell people to take interest if the email matches some criteria. So it isn't the intrusive one, but it could automate finding people to intrude on.
Do you think that this hasn't been happening? As someone posting on tech site, you should be aware that they scan emails for keywords, probably google and certainly the NSA.
Furthermore, training LSTM networks on emails has nothing to do with what you are concerned about.
The worrisome thing is that it while it used to be the privileged who insulated themselves from the "real" world by having staff answer their email, this would let everyone do that, and that might give everyone more tunnel vision.