We can only hope for the day when NYT comments are automatically-generated.
The clichés of business email speak, which this article says are “recognized for their usefulness”, the article shows are actually rather useless—a point of friction in the communication process that could be eliminated with no loss of communication, content or real information. Similarly, this idea that we are “professional emailers” because we spend a third of our office time emailing regardless of our specific job, I think has the pretense of profundity, but no real significance. It’s like saying that 100 years ago, most white collar workers were professional hand-writers because they universally spent a substantial portion of their time in their offices handwriting. Email is just a format for communication. Indeed, office workers spend much of their time communicating, and none of their time actually making stuff. But I loved the idea of an infinite regress or chain of email communications generated by “AI” Smart Compose and Smart Reply—basically bots talking to each other. What this really shows is how far behind “AI” is from the capabilities that we imagine for AI.
Smart Compose is great. I'm the sole assistant to a freelancer, and I often have an incredibly heavy workload. I handle all communication across several time zones, and a typical work day is 12 hours long. I don't usually use Smart Reply, because I don't like to sound so casual. But Smart Compose has saved me so much time and allows me to accomplish more.
Using this technology doesn't mean I'm lazy and it doesn't prevent me from composing original replies. I suppose its usefulness depends on what your business uses email to accomplish. Email is the primary mode of communication in my sector of the fashion industry, as we all work internationally, are OOO/working remotely/traveling/on a shoot/etc.
My friends and I have to work from form letters for many parts of our jobs, and predictive text technologies can help us return to our lives faster following a work-related interruption. I had a friend over for drinks one recent Friday evening. At 9pm she received a call sheet she needed to send out to her artists. She has done this from her phone so many times that it knew exactly what she wanted to type. The call sheet was sent and we were back to catching up within 90 seconds instead of 10 minutes. It's a far cry from the mythical 3-day work week, but in my opinion, anything that helps relieve the pressure of a demanding job is worthwhile.
1
Keynes' prediction was way off. The savings from automation led to fill the freed up time with other work or more work. What before was done by an assistant, I have to do it now and spent inordinate hours answering important and run of the mill correspondence and calls in the form of emails. Same with voicemail. It makes it so difficult to finally establish communication since both parties postpone answering the phone at their convenience.
There is never savings on the employee's side. The employer will own the savings by any step in automation.
4
I hate the new "smart reply" and "smart compose" in Gmail. I only read this article in hopes that you might've mentioned how to turn them off!
2
@imandavis: For me the tool offered to turn it off, which of course, I promptly clicked. But if you missed it by any chance, go to settings by clicking the gear icon on the top right hand side
Go to Settings -> In General tab set smart replies to off and writing suggestions to off.
This should do the trick
1
Great - so I will get more meaningless e-mails that just tell me "I will get back to you". The biggest challenge of today is not the numbers of e-mails we have to write, it is the number of e-mail we receive. So, dear Google, please shift your priorities from AI-supported writing to AI-supported digesting of emails.
2
If we're too lazy to type "Let me know what you think" when we're e-mailing from our computers, it's time for us all to lay down and die.
4
I prefer to do my own thinking. I will not subcontract cognitive functions to AI, auto-correct, auto-fill forms, etc. Call me a doomsdayer if you must. We are slipping and sliding ever closer to a Matrix-like future.
3
i was confronted by one of these (words we don't use in this civilized forum) things the other day. Took me forever to get around it as it presumed something about what I was doing that had no relation to what I was doing, and would have been none of its (see above) business in any case. Just makes me less trusting of the technology I use.
2
The author's last paragraph is telling, in that it points to the recipients of these automation benefits. One only has to look at our level of inequality to see who it belongs to. As more and more labor can be automated and improved ownership of the technology becomes critical, and in our country, it is owned by a handful of technology magnates, the members of their management circle, and the wealthy that own those companies. The technology itself can be wonderful and useful, but the lopsided power imbalance of ownership means that technology and corporations can comfortably and easily rid people of the work lives and livelihood.
2
I've worked in technology for 30 years and as a software developer for the past 15, and I was always been aware of the 'crushing wave' which I needed to stay ahead of. Although software developers are often seen in demand, that will not always be so. Software, by its nature of relying on rules, patterns, and practice, will automate more and more of itself. New technology often automates older work, and one way of dealing with that is to automate the task/role before you/it are crushed by the next wave. Essentially, you ride the new wave that is going to crush the results of the old wave. Not to minimize the concern, but workers not accustomed to, nor directly involved in, automation can feel threatened by the loss, but it is a threat technology personnel live with continually.
1
The basic truth of automation is that when you people use rules and principles to make decisions, those same rules and principles can be codified and automated. Regarding Google's suggestions, they are rarely sufficient to express what I need to say, nor are they effusive and detailed enough for my personal communications, since most of the offered responses are little more than the equivalent of a grunt. That said, most people might likely find the reflexive responses more than adequate. Even though people are not always conscious of the rules they follow in communication, they exist, and as such can be divined.
1
Dont get me wrong i love AI and deal with it to a degree but this sounds like in another ten years AI will not just be helping out but directly emailing each other...
3
"They surveyed “highly educated baby boomer or Gen X” subjects who were mostly “managers or professionals” working in office jobs and found that they spend on average a full third of their workweeks “processing” email. "
Or you could say the managers and professionals spend a third of their workweek engaging in correspondence with company staff, customers/clients and suppliers.
Canned responses and letter templates have been around in various forms for a long time and people will recognize them.
4
How about automatically responding to all emails with, "dope."
I love that term. It's so appropriate for so much nowadays.
5
Ah the erudite contributors and commenters of the New York Times. They certainly know what people do to get through the day.
People who do tech support know other things. They know how people get into ridiculous routines because, well, they are too dumb to figure out how to do it better. To forward an e-mail, they print it, scan it in, and then send it as an attachment. To use a spreadsheet they print it, add, subtract and multiply using a calculator as needed, fill in the results by hand on the printout, and enter them back in the original.
Clippy's problem was that not enough "average" people were using PC's at the time. Believe me, that isn't true now. Yeah, saying "you appear to be...." to a tech writer is ridiculous. Some users though, they need all the help they can get. Even if it is an animated paperclip.
@Daedalus - I'd wonder what companies you have worked with since I have never met anyone so ignorant as you describe. Even then, when there are stories, the problem seems to be caused by limits of users' technology - deleting system files because there isn't enough disk space - or the limited social abilities of their tech support - they did not communicate and train sufficiently.
4
How do I automatically open this story, sent me via Google email, in your app? Instead of my Safari or Google browser?
I am in the office today and will not be able to respond to your email.
9
Spellcheck is useful. "Your order is on its way" is a useful notification. But A.I. packaged sentiment from Google's Clichés R Us enterprise in their Whatever server farm is not human communication. It's 20,000 variations of Orwell's "double plus ungood." It's automated noise. What's the point if people tune it out?
These programs are attempting to read my mind. STOP IT!
I paid a substantial amount of money for a Samsung S9, running Googles licencensed Oreo android operating system. Googles gmail notifications are spotty at best...Time and time again I've missed reading messages in a timely fashion. At times, by days. No one at Google or Samsung is talking when I ask for help. Yet...the AI robots won't shut up.
This is something consumers don't need. It's more information gathering on Google's part, at the expense of consumers having Google document every facet of communications, whether it voice line, email, or text messages.
1
Easy-peasy - Turn it off. I did. It was so annoying.
3
It can't be a good idea to use the cache of some 500,000 emails collected during the discovery phase of the Enron investigation to train email to provide smart replies. Wouldn't that teach AI to respond in ways that are in concert with illegal activity? Sounds like a bad plan.
6
Better question: Will someone program an AI that smacks email-using businesses when they EVER compose as HTML?
All rich-HTML-formatted (as opposed to text-only) emails do is give security breachers a ridiculously easy way to do their work. The EFAIL exploit and of course the DNC attack[1] are but two massively painful examples that we'll continue to pay for, for a long time.
They also make it easy to hide creepy trackers like pixels and scripts. Unfortunately, that's probably EXACTLY why businesses use them even more. Even NYT's comment-published notifications are so formatted, when they used to be simple text.
Sigh.
[1] https://www.nytimes.com/2016/12/13/us/politics/russia-hack-election-dnc.html
1
This sort of artificial intelligence might be useful to newbies or people who lack basic grammar skills. The thought that email providers use this technology to predict our thoughts as we type by offering alternatives in terms of composition along with one-click auto-response options is another step closer to society becoming brainless robots.
1
Anyone who really WRITES a mail can not use these. They are great for replacing the people who needed to be replaced.
2
Carry it to its logical extreme: we won’t have to think anymore, nor will we need to know how to write. A few stock phrases and—voila—we’re done. It makes social relationships so much easier. We don’t even have to have them. Gosh. Thanks, engineers!
3
This was a very interesting article, and I really enjoyed it a lot! John Hermanis did a good job explaining all the technologys for Google and smartphones."Depending on what your current inbox looks like, this might not require much imagination at all." In conclusion thank you.
I like writing.
I like thinking.
I like composing my own sentences.
I'll stick to writing my own emails.
1
All I want to know about this is how to turn it off, please.
3
Already our country is filled with people who say everything is “awesome” in a way that is not an expression of thought but a substitution for it. Now Google is providing a whole universe of responses that work the same way.
Awesome.
4
Things like spell check and word suggestion are distractions that invite the writer to back up and correct, or to stop and consider whether what is suggested will correctly represent the thought. I'm tempted to turn it all off, but I'm afraid I would forget to turn these on after composing prior to sending. I wish there was a button for such things near the send button, instead of them being live throughout my composition.
Much like with texting, in email, I tend to drop the salutation, and my name automatically appears as a signature, so I usually only write my thoughts and click send.
I already realize there is no such thing as privacy on the internet, but for me this type of AI is overstepping that boundary. I will answer my own emails thank you with my own responses. This type of automation only furthers the dumbing-down of a society unable to articulate in there own thoughts and words. Could not find out how to turn off this “feature” on my gmail account fast enough.
This appears to be a virtual assistant aimed at guessing what you might want to say, then preparing draft replies and asking for your approval/edits to their draft.
I'd much rather have a virtual assistant that (based on their vast store of knowledge) can suggest better/more effective word formations for what I am trying to communicate.
Ok so that’s what machine-to-machine communication means. I’ve always wondered.
Has the potential to be incredibly time-saving and liberating for employees, until the company goes bust because nothing actually gets done.
5
I am a very happy user of an app called Text Expander that has allowed me to create a pool of form letters that I can activate in an email message by typing a brief abbreviation. The form letters can include prompts for locations where I want to personalize the message, by inserting a name in a salutation, for example. I receive inquiries, applications, deposits, etc. for a program I direct, and Text Expander has vastly streamlined my work. The app has many other bells and whistles, and I strongly recommend it.
1
Smart Compose often uses hackneyed phrases. Sometimes these are good enough and sometimes they remind me to be more creative. Seldom do the Smart Reply phrases fit my own style, but they are handy once in a while when I am busy.
3
On the plus side, I am using Gmail's prompts to learn to write more creatively. I ignore what Gmail suggests and look for a more mindful way to communicate.
15
Great. Now automated systems can generate more email, and glibly answer them, while just getting a real answer from a real person remains nearly impossible.
14
So like (and I am 50) no one in any modern hyperscale organization actually uses email anymore. So passe. If your peeps are using and depending on email for productivity then you really need to reassess your IT infrastructure. Many of the newer platforms require less not more actual outdated communication like email anymore and many of them like Slack/Github/Atlassian are open source or very very cheap to use.
4
@Miguel
In my world, email is now what the formal business letter used to be. Email indicates a transaction that demands serious documentation and response. Everything else is texting.
11
@Madeline Conant If you work in an industry (securities and investing, for example) where regulators require continuous archiving of all communication (both internal and external), email is the only way to go.
1
It remind me, I was into computers when I was a teen in the 80s and my grandfather said to me once "computers may be smart, but they'll never write a symphony."
And in my youthful arrogance I thought to myself "well neither will you."
A frequent debate I have with a friend is over whether or not AI is actually possible and I like to tell him that he gets too caught up on something that acts like a human when AI that is as "smart" as much less intelligent lifeforms has it's own uses and dangers too. An AI that is as smart as a bumblebee could have all sorts of applications and risks. Pigeons have been trained to do "jobs" that at first glance might require a human and I'm sure we are indeed quite close to having AI that is as smart as a bird.
My general point being that there is a whole lot of stuff (both good and wicked) that humans do that doesn't require any particular degree of actual intelligence to accomplish.
4
In the late 1950s Senator Thomas E. Martin of Iowa introduced prepared on small credit card size metal cards paragraphs that the electric typewriter would rapidly type. The secretaries had references books for all constituents who had written and would read the letter, type some personal remarks such as names and anything he wanted added, and then type the prepared responses to all the standard inquiries.
As a senator he had the third highest mail yet his office budget was based upon his state’s population! Iowa had the highest adult literacy in the nation and they wrote.
His last year as a senator, 1960, more than 100 senators asked him and his staff’s help in automating the routine while keeping the personal touch with the remarks and questions asked in their replies.
It would be nice to have software that could on an adjacent screen have available relevant text.
6
Very interesting article! Lots of food for thought. I'll get back to you soon. Thanks for all your hard work.
120
@Jason,
Very funny!
1
I loathe those prompts that have started appearing in gmail. I can imagine a place for them, in large scale, largely impersonal interactions such as some Customer Service engagements. However, I hope the day never comes when my interactions with friends and co-workers are routinely reduced to canned responses.
40
@mosselyn, you can turn the feature off. Google how to do it.
@mosselyn It's the same in Outlook.
“Sounds great!” “Very cool!” “Love it!” ...this is the extent of supposed AI? Laughable. Email, for better or worse, in still an integral and complex part of most everyone's life. Most meaningful email exchanges (and conversations by extension), like language itself, is very complex in meaning, tone, and implication. Many email exchanges involve complex decision making, judgement, and direction setting that AI has no clue about. Today's 'artificial intelligence' is just that (i.e. artificial). If thinking that Google can help you by saying, “Sounds great!” “Very cool!” “Love it!” ... then chances are you don't have to say anything at all.
28
@JeffB: Are all of the emails that you receive thought provoking? Do all of them require a considered response? It may be less that you don't have anything to say and more that you don't have to say anything.
1
Can my AI talk to your AI while I go for coffee?
41
@Ben
Actually, neither of us is needed at all ... the two AIs can handle everything!
1
The laziness of people today astounds me.
28
I take great care composing email messages to business associates. I use my PC and try to cover every detail. In many cases, it is obvious that the very short reply was written on a smart phone during a meeting without full attention. Three or four volleys must follow to correctly address the matter. In my opinion, AI will only make this worse.
31
@Richard
Totally agree. People do not read! They skim--or stop after the first sentence.
I always scale up my font in an email to a slightly larger size with slightly more than one space (1.25) between lines to make it more readable (and increase the chance that it might actually be read...)
If I need answers to more than one question, I have taken to using bullets. I know someone else who sends a new email message for each question--which can become extremely annoying, especially when all the questions concern the same topic/project.
And I agree that AI is indeed going to exacerbate our lack of communication skills.
@Richard Exactly. Next someone will try to come up with AI to do our reading for us. (In the meantime, seems like many substitute videos for reading.)
Given recent privacy issues, I wonder where the AI is applied. Is it applied on the user's device or is it applied on Google servers?
If it is applied on Google servers then presumably Google is receiving in analyzing everything that the user has received or is typing.
Serious privacy issues may abound.
34
@Biff Trout that has always been the deal with GMail. Initially it was to offer you "more relevant ads", but now every free tool online is about data collection. There is no such thing as privacy online -- the moment you connect to another server, someone is capturing and storing your data.
8
@Biff Trout - Google is way ahead of you. Their algorithms have been reading our emails since 2012. Google announced last July that it would stop using consumer Gmail content for ad personalization, but their current privacy policy still states, “We collect information about the services that you use and how you use them. […] This includes information like your usage data and preferences, Gmail messages, G profile, photos, videos, browsing history, map searches, docs, or other Google-hosted content. Our automated systems analyze this information as it is sent and received and when it is stored.”