June 25th, 2010
|01:16 am - Talking With Computers & The Future of Work|
I this recent post, I mentioned the idea that call centers and other front-end customer service jobs would soon be automated, and after reading this article it was clear that soon was closer than I expected While still in testing at Microsoft, we already have programs like this:
The Digital Assistant That all looks impressively useful, and it looks fairly clear that in 5 years digital assistants will be beginning to achieve widespread use, and in a decade they'll be ubiquitous. I personally find this fairly exciting, both because it's fascinating technology and because I often find talking to strangers on the phone and so love good automated systems. I was also very interested in this bit that immediately followed the description of the digital assistant:
“Hi, are you looking for Eric?” asks the receptionist outside the office of Eric Horvitz at Microsoft.
This assistant is an avatar, a time manager for office workers. Behind the female face on the screen is an arsenal of computing technology including speech understanding, image recognition and machine learning. The digital assistant taps databases that include the boss’s calendar of meetings and appointments going back years, and his work patterns. Its software monitors his phone calls by length, person spoken to, time of day and day of the week. It also tracks his location and computer use by applications used — e-mail, writing documents, browsing the Web — for how long and time of day.
When a colleague asks when Mr. Horvitz’s meeting or phone call may be over, the avatar reviews that data looking for patterns — for example, how long have calls to this person typically lasted, at similar times of day and days of the week, when Mr. Horvitz was also browsing the Web while talking? “He should be free in five or six minutes,” the avatar decides.
The avatar has a database of all the boss’s colleagues at work and relationships, from research team members to senior management, and it can schedule meetings. Mr. Horvitz has given the avatar rules for the kinds of meetings that are more and less interruptible. A session with a research peer, requiring deep concentration, may be scored as less interruptible than a meeting with a senior executive. “It’s O.K. to interrupt him,” the assistant tells a visitor. “Just go in.”
As part of the project, the researchers plan to program the avatar to engage in “work-related chitchat” with colleagues who are waiting. Software like this also make me even more certain that if actual conscious AI is possible (which is at this point completely unknown) it's likely to be something that arrives in a series of gradual steps, starting with things like this and (if it indeed turns out to be possible) a couple of decades later eventually ends with humans talking to electronic people.
The conversation could be about the boss’s day: “Eric’s been in back-to-back meetings this afternoon. But he’s looking forward to seeing you.” Or work done with the boss: “Yes, you were in the big quarterly review with Eric last month.” Or even a local team: “How about that Mariners game last night?”
In any case, this is all impressive, incredible, and exciting, but also makes it quite clear that we are going to need something like a guaranteed minimum income, as well as a changing definition of work. By 2020, many call center, low level tech support, and receptionist jobs, as well as a number of transcription jobs (which according to the article are already being automated) either won't exist or will be in the process of being phased out. This is a process that seems to start with outsourcing and ends with automation - a process we've already seen in a great deal of manufacturing work. One obvious question is what will most first world residents do for a living in 2030? I don't have an answer to that question, and I doubt anyone else does either - all we have is some guesses. It looks to me that we are beginning a revolution in office work exactly comparable to the revolution in manufacturing work that started 40 or so years ago when industrial automation started to become popular.
Current Mood: thoughtful
|Date:||June 25th, 2010 10:06 am (UTC)|| |
I wish I disagreed, but I mostly don't. We desperately need some sort of GMI, but I'm not at all certain the US will get anything like that. I expect most of the first world to go that way within 30 years, and likely in well less than that, and like sane healthcare, I expect the US to lag badly behind. If that happens, then it will definitely be time to look into how to go about moving to Canada or the UK (having one partner who recently found out that she had dual Canadian citizenship made me very happy indeed). I'm betting that in 15-20 years most of the first world is going to look very nice indeed and the US will suck even more for anyone who isn't rich.
The ever-present question about the US is exactly how bad will things need to get for actual improvements to take place. Things had to get pretty darn dire for the New Deal to become possible, and I worry that something similar will need to happen for the next round of major changes to occur. I'm hoping that the tea-baggers do enough damage to the Republican party to at least make such changes a bit easier.
Edited at 2010-06-25 10:08 am (UTC)
|Date:||June 25th, 2010 11:19 am (UTC)|| |
That's something SF always seems to miss about the future: yes, many technological improvements will come, but they won't create a rising tide that floats all boats.
Serious, sweeping change would be nice, but the powers-that-be are too good at taking away the option of real change.
|Date:||June 25th, 2010 11:15 am (UTC)|| |
I've been trying to think of jobs that can't be done automatically or outsourced. Seems like skilled trades -- plumbing, car mechanic, that sort of thing -- are it.
Creative professions, live performers and personal service will also have their place.
Especially if automated serive end up the norm in many low end sit-down restaurants, going to a place with all human wait-staff will likely be something people will pay for. Similarly, I expect that there will continue to be work for barbers, people giving massages, and various other spa and personal care jobs that seem like people would strongly resist using automated versions of. I'm also not expecting doctors or nurses to be automated anytime soon, but a number of the more purely technical aspects of medicine may well be.
|Date:||June 25th, 2010 02:06 pm (UTC)|| |
I don't disagree with the facts either..
but the facts really terrify me in this case. I think a GMI like you suggest is the best way to go but I think that what will end up happening is a higher threshold for systemic unemployment.
As someone who spent a number of years working technical support and did the transition from tech support to tier 2 to back end work my immediate concern is that there won't be a real career path that someone can follow to get into IT(although this concern can be extended to other professions that need entry level jobs that a person advances through).
Where I worked it was routine to see people who were just out of school, recovered addicts, people who had escaped abusive relationships and other sorts that were basically trying to get their life back together. The downside of the automation trend is that the jobs that people use to get back on their feet are the very ones that are automated. Some of them will just suffer quietly. But some of them will get very cranky.
I'm skeptical about the time frame for this. The AI receptionist actually sounds a bit plausible, because it doesn't have to be general-purpose. It has a pretty limited domain of things it has to think sensibly about. But I don't think that that necessarily indicates much about the time-frame for strong AI, because generalizing intelligence has always been an enormous difficulty factor for AI work. So I think that the time-frame is kind of extended, that we shouldn't expect too much of it. Similar for transcription: the bigger the expected vocabulary is, the worse computerized transcription gets - and it still needs human checking a lot of the time, because of the variation range of human speech. So that one, too, I don't think is on track to arrive soon.
Strong AI is a big unknown. If it's possible (and I think it is), it will drastically change society, but it isn't necessary for much of this to be realized.
It sounds like computer transcription is doing remarkably well, and from what I saw in a few articles I've read, in a decade I'm not certain that the error rate will be much higher than human transcription. However, the big deals are receptionist and call center automation , since these will affect far more jobs.