end of email as we know it

[Bongo]

AI used to check spelling in my emails, and I really appreciated that.

Then AI started correcting my grammar,and I found that helpful.

Then I would type the first word or two of a commonly used phrase, and AI would complete the sentence for me.

Then AI started finishing more nuanced sentences based on the email's context, and to my surprise, it was right most of the time.

The saving grace was/is that I had the option to use the "hints" or not.

Today I had a conversation using email, and AI summarized the previous message. I didn't ask for it, AI just did it on its own, but the humour and sarcasm were lost and a totally different impression was conveyed. It was just a silly back-and-forth I had with a friend -- but what if AI starts summarizing all messages, and we become lazy or dependent on the summaries? And the summations are longer, more detailed and nuanced...

SO... a friend completed a 1,000-piece puzzle in a few hours and I jokingly accused her of cheating by turning the pieces upside down so that the picture wouldn't confuse her. She played along and asked for more cheating tips, so it's kind of obvious I was joking, but what if the subject was about something other than finishing a puzzle?

 

[brianvds]

Me, I absolutely detest it when machines think for me. The very worst is autocorrect, which I tried and failed to turn off on my tablet, so now every time I WhatsApp, I have to constantly go back into messages to correct autocorrect's blunders.

 

[stlukesguild]

Yes… autocorrect often “corrects” vocabulary that is correct but not recognized by the program… especially terms related to art (or other disciplines), non-English words, and proper names.

 

[Enyaw]

Makes say things I never intendo.

 

[Jo Castillo]

My daughter says AI guesses from what it knows for sure. It gave her the wrong address and open times of a restaurant that actually had a website with correct info. Guess it took the average of restaurants in the area? Makes me wonder how well it will work in the medical field, or guiding airplanes, cars, etc. The self driving cars in Austin, TX., have been passing school busses that are stopped with red lights. However they found that more people had passed than the self driving cars. Ha.

 

[brianvds]

Yes… autocorrect often “corrects” vocabulary that is correct but not recognized by the program… especially terms related to art (or other disciplines), non-English words, and proper names.

And there seems to be no way, at least on my Android device, to turn it off. So now I have a constant battle against it.

 

[Bongo]

Also turns out the "sarcastic joke" I made in the email: accusing my friend of cheating by turning puzzle pieces upside down so not to be confused by the picture ---well---the joke is on me because--
-
Little did I know that nowadays they PRINT LETTERS on the backs of the puzzle pieces. So you turn all the pieces upside down then sort them into A,B,C,D... groups, making sure the letters are also in the proper orientation - not sideways or upside down. Then turn the pieces picture side up and fit the groups one at a time. This obviously greatly simplifies working the puzzle.

So, irony of irony, the AI email overview "thought" I was serious - because nowadays turning the puzzle upside down IS how you "cheat". So what was lost in the translation was not my humour but my stupidity.