Great, thanks! Thank you for your help! Nope, I’m good. Thanks, you too! You too! Same to you! Great, thanks for letting me know! Looks great! Not sure yet. No, it’s not. Awesome, thanks! Looking forward to it! Will do.
My AI just doesn’t get me.
The Google email service has an artificial-intelligence feature that apparently has been around for a year or so but, for whatever reason, has only recently inserted itself, uninvited, into my life, like that U2 album that’s on your iPhone for no reason. You probably are familiar with this email feature already: It suggests colorless responses to certain emails, so that if your answer is going to be “Thanks, awesome!” then you don’t have to go through all the trouble, stress, and privation of typing out 13 letters and two punctuation marks. Just click on the answer you want.
Easy to mock, though I suppose it does serve some function: If you do a great deal of your text-based communication on your phone (perhaps while sitting in the middle seat in economy on a flight to Columbus, Ohio), then it might be easier to point and press than to type. Maybe the suggestion isn’t exactly how you’d write the response, but it’s good enough.
We used to think AI was going to take over the world and declare war on the human race. Instead, it has taken over my email account and declared war on my correspondence: I do not think that I have often — if ever — written the words “Awesome, thanks!”
Correction: I did not think that.
Because email lives forever, I now know that I did write almost that (mine was “Awesome” period, capital-T “Thanks,” exclamation point) to Jay Nordlinger last winter. Jay had forwarded to me an email from the musicologist Robert Marshall. On the subject of the politics of great artists, I had written: “Who knows what Bach or Bernini thought about tax policy?” Robert Marshall knows what Bach thought about taxes, as it turns out. Bach wrote a letter to his cousin complaining about the duties he owed on a gift of wine sent to him, taxes that came to about $20 a quart. Marshall wrote eight paragraphs, beginning with: “On November 2, 1748, he wrote a letter to his cousin, Johan Elias Bach, who lived in the Bavarian town of Schweinfurth, acknowledging that he had received his gift of an ‘excellent little cask of wine, for which I send you herewith the gratitude I owe you.’”
And I replied: “Awesome. Thanks!”
Which is, in retrospect, pretty lame, even if it was my “Awesome. Thanks!” and not Google’s.
Reihan Salam, I now know, is an occasional “Awesome, thanks”-er. But I think that seems more natural to the younger man: He’s practically a Millennial.
But, wait — did Reihan really write that? Maybe he’s not an “Awesome, thanks”-er at all. Maybe I’m not even getting emails from him, but from his electronic monkey-butler. And if I use my electronic monkey-butler to answer his electronic monkey-butler, and it answers back, and so on, at what point do Reihan and I stop being part of the conversation? Reihan says he always adds something to the message, “if only to defy the suggested reply.” I hope he doesn’t give in.
Because I am a writer, it wounds my vanity a little bit that the nerds in Silicon Valley think that my emails can be composed by a robot. I put thought into those emails. (Sometimes.) We creative types are supposed to be occupying the last defensible position against automation. We willingly accept all sorts of occupational hazards — deadlines, the occasional barbaric editor, and, in the newspaper business, the guy who shows up at your office with an advanced case of rage-face because his name is in the police blotter for the DUI he got on Saturday night — but automation has never been a worry, because we do not think that you can automate what we do.
But a lot of people think that, and many of them are wrong.
You can, for example, automate a lot of what lawyers do. LawGeex put its AI lawyer up against a team of big-shot human lawyers (representing big companies such as Goldman Sachs and fancy law firms such as Alston & Bird) and asked them to review five nondisclosure agreements, which are common as dirt in business transactions. The AI stomped the meat. The AI was 94 percent accurate in its work, as opposed to 85 percent for the human lawyers. To produce those inferior results, the human lawyers took between 51 and 156 minutes each to complete the task.
The AI took 26 seconds.
I work in a time-sensitive business, and I am not always what you’d call strictly right on time. (Ask Reihan.) The shadow falls, and deepens.
That’s life in the 21st century: Of course your email is reading your email, which is weird to think about, and I kind of would feel sorry for it, if it had feelings or got bored. I once read a few despairing tweets from a poor little gnome at Media Matters who had been assigned to the Kevin D. Williamson file, meaning that she had to read everything I wrote and look for things that could be broadcast under the headline “Evil Conservatives Are Evil!” I felt sorry for her; I don’t even want to read everything I write, and I write a lot. Worse, I don’t think my email is the most exciting — here is a selection of recent subject lines:
FASB and GASB to Cohost IN FOCUS: Not-for-Profit and Governmental Accounting Webcast for Academics
Cato’s Fintech Summit is this Wednesday!
Copy deadline: October 10
Error in Burt Reynolds column
Grange urges Congress to act on looming Part D cliff
Not so exciting.
We live in a state of surveillance. Your email is reading your email, and your car is watching you drive. For years, lane cameras have been alerting drivers if they drift into another lane, but that’s just watching the road — now they watch you. From David Szondy at New Atlas:
Volvo’s approach relies on sensors installed in the dashboard. Called Driver State Estimation, this consists of small LED lamps that shine invisible infrared light on the driver’s face. Sensors pick up the reflected light, and the system uses face recognition technology to determine if the driver is awake and alert by measuring such factors as how wide open the eyes are, along with the position and angle of the head. Volvo also sees this face recognition technology as a way to unobtrusively personalize the car by letting it automatically recognize the driver and adjust the seat and cockpit settings, as well as using face tracking to adjust interior and exterior lighting based on which way the driver is looking.
Of course, it’ll parallel park for you, and tap the brakes if needed, and do all sorts of other things. They’ll be truly autonomous eventually and, after proving that AI drivers outperform human drivers as much as AI lawyers outperform human lawyers, the trip from “Permitted under heavy regulation” to “Mandatory” will be a short one, involving no driver fatigue at all.
Our world is full of wonders, and no sensible person would trade a 2018 standard of living for a 1958 standard of living. But there is something lost. Harley-Davidson is bringing out an electric motorcycle, which is great, but one of the selling points is: Young people don’t know how to work a clutch or shift gears. (Ronald Reagan, out and about on the ranch in an old Jeep, once asked his Secret Service contingent whether one of them would mind driving him back to the house. None of them could, because none of them could drive a stick.) Progress comes and takes its toll, and the world grows more strange. Of course your car is staring you in the face, wondering how you’re feeling, whether you need a nap. Of course Google can write that email for you. And that’s all great. But even a technophile such as myself must sometimes wonder what the point of all this really is — and whether it is wondering the same thing about us.