Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
AI Robotics

When AI Can't Replace a Worker, It Watches Them Instead (wired.com) 50

Whether software that digitizes manual labor makes workers frowny or smiley will come down to how employers choose to use it. From a report: When Tony Huffman stepped away from the production line at the Denso auto part factory in Battle Creek, Michigan, to talk with WIRED earlier this month, the workers he supervised were still being watched -- but not by a human. A camera over each station captured workers' movements as they assembled parts for auto heat-management systems. The video was piped into machine-learning software made by a startup called Drishti, which watched workers' movements and calculated how long each person took to complete their work. [...] Denso's use of Drishti shows how some jobs will be transformed by artificial intelligence even when they're unlikely to be eliminated by AI anytime soon. Many jobs in manufacturing require dexterity and resourcefulness, for example, in ways that robots and software still can't match. But advances in AI and sensors are providing new ways to digitize manual labor. That gives managers new insights -- and potentially leverage -- on workers.

Some workers say the results are unpleasant. Last year, Amazon warehouse employees in Minnesota staged a walkout to protest how the company uses inventory and worker-tracking technology. They allege that Amazon uses it to enforce a punishing working pace that causes injuries. The company has disputed those claims, saying it coaches employees on how to safely meet quotas. Workers at Denso were initially wary of the prospect of being video-recorded all day to feed machine-learning algorithms, but Huffman says they have since come to appreciate Drishti's technology. After something goes wrong, workers can now look at the data and video with their managers, instead of having to hope bosses take their account of what happened seriously. Huffman says having a constant readout on productivity also helps managers be more responsive to nascent problems. "If somebody's struggling, not every associate is going to call for help," he says. "If we see their cycle time is jumping through the roof, we can go over and say 'Are you having any issues?'"

This discussion has been archived. No new comments can be posted.

When AI Can't Replace a Worker, It Watches Them Instead

Comments Filter:
  • by ludux ( 6308946 ) on Saturday February 29, 2020 @12:20AM (#59780944)
    Put the cameras on the managers and calculate how long each manager takes to complete their work, so it can helpfully it coach the managers on how to safely meet quotas. If something goes wrong, workers can now look at the data and video of their managers. Having a constant readout on manager productivity should be a great thing, after all, if a manager is struggling, not every manager is going to call for help. Methinks this would get shut down right quick if the camera were pointing at the bosses rather than the people serving them.
    • by waspleg ( 316038 )

      Yes, the "but it helps us help you!" implication made my stomach turn. This is going to end very poorly. I'd rather be a battery in The Matrix than live this Black Mirror episode.

    • by Anonymous Coward

      The flaw in your argument is that most workers actually do something of value, which is measurable.

      However most managers do nothing of value, which obviously does not require measuring.

      • I’d like to quantify exactly how much nothing they’re accomplishing though. I’ve known a few that probably do tens times as much nothing in a day as others do all week!

        Really though managers that do a lot of nothing are pretty decent. It’s the ones that try to do something and make a giant mess that are a problem.
      • What are you talking about? Managers go to, or conduct, tons of meetings! When they aren't doing that, they conduct training classes for peons. These things are easy to measure, they are all exactly one hour each, even if the content requires only 5 minutes to explain.

    • by 93 Escort Wagon ( 326346 ) on Saturday February 29, 2020 @01:51AM (#59781082)

      Put the cameras on the managers and calculate how long each manager takes to complete their work, so it can helpfully it coach the managers on how to safely meet quotas.

      What work is it you think most managers actually do?

      The good managers I've had basically got out of my way and let me work. I have no idea what they did, most of the time, other than to facilitate initial meetings with clients. That's useful, but hardly a full time job.

      The bad managers I've had have all tried to involve themselves in processes they didn't understand even partially, which generally pushed the completion time of the projects out 200-300%.

      • by Dutch Gun ( 899105 ) on Saturday February 29, 2020 @02:03AM (#59781108)

        Like you said, the best thing your manager can do is to stay out of your way and let you get your job done. But another important job is keeping their managers from breathing down your neck with whatever sort of idiotic schemes they inevitably come up with - or at least trying to minimize the pain those idiotic schemes cause.

        • by JaredOfEuropa ( 526365 ) on Saturday February 29, 2020 @06:19AM (#59781350) Journal
          So... if their most important contribution is to not interfere with our work and keep other managers from interfering, why do we need these managers at all? Sounds to me we would be better off firing them.

          Kidding, of course. A lot of what managers do is useful or even vital: planning, strategy, conflict resolution, resource management, prioritizing, budgeting, administration, decision making, and so on. And not much of that lends itself well to analysis by an AI using a camera. But regarding their "idiotic schemes", what surprises me is how terrible most organisations are at measuring the effectiveness of their management. They look at a subset of aspects, like the output of their department, but at best those metrics are only part of the picture, and they are often more a result of the quality of the team than of the individual manager. And even organisations that can kinda sorta spot good managers and bad ones, they still fail at seeing what exactly makes those managers good, and try and replicate that in the rest of the organisation (and more importantly, culling bad practises and bad managers).

          So much effort is spent on executive search, management development programmes, assessments, and so on, but I've not seen an organisation able to consistently hire or promote competent managers. Who knows, maybe some day an "AI" will be able to help with that too.
      • by Kjella ( 173770 ) on Saturday February 29, 2020 @06:33AM (#59781352) Homepage

        What work is it you think most managers actually do? The good managers I've had basically got out of my way and let me work. I have no idea what they did, most of the time, other than to facilitate initial meetings with clients. That's useful, but hardly a full time job.

        The irony is that this is exactly what they think about good developers. Their systems keeps chugging along without hardly any downtime or significant bugs. "I have no idea what they did, most of the time, other than to solve a few minor issues. That's useful, but hardly a full time job."

        The bad managers I've had have all tried to involve themselves in processes they didn't understand even partially, which generally pushed the completion time of the projects out 200-300%.

        Yeah that's a manager that fails at his other duties, like negotiating reasonable expectations, getting stable priorities, protecting the team from distractions and disruptions and so on. They just see the team fail to meet milestones and deliverables, get yelled at his bosses to either start meeting goals to or to be fired.
        They don't see any other option than to try cracking the whip to make people work faster and pile on tasks with increasingly impossible deadlines and panicked re-plannings that only causes more chaos.

        You don't see the work good managers do because he, like you, is working on keeping things under control. If you constantly make small course corrections, if you constantly nip problems in the bud, if you fight back on any vague specifications, scope creep, unreasonable expectations and lead times then they settle it and the developers are freed up to be the "doers" they're supposed to be. Like you he's fixing problems before they become problems, if you think management is easy try it. You try to keep the shit from rolling downhill on your team and see how easy it is.

        • The problem is that upper management is not subject to these pressures to perform. They advanced on the Peter Principle (you advance to your highest level of incompetence), and are above the fray to digitize every aspect of justifying your pay. AI can be fooled too. In the old Soviet days, the workers had the system down to a science. They performed at a level just acceptable to the humans that were managing them, with a collective effort to keep productivity growth as minimal as possible to keep higher
        • by scamper_22 ( 1073470 ) on Saturday February 29, 2020 @01:56PM (#59782306)

          Every one exists within an organization.
          Eventually, you should learn to do one of 3 things.

          1. Operate within that organization
          2. Change the organization
          3. Leave that organization

          I was like most software engineers/developers when I graduated. I spent most of my career in R&D where in general I believed creating a good product and doing a good job was how the organized R&D. In general, I'd say it was true. Obviously every job has politics, but in general it was true. I largely operated within the organization by doing a good job and building good products.

          Later in my career, I joined a bank. At first I tried to operate as I had always done. Doing a good job and building great products. I still dead well at the bank, but I noticed that wasn't how things were organized. Project managers, contractors, project charge codes, funding... were all very much a part of the organization. I'm not about to change how the bank operates, so I simply adapted to the organization. I charge my time accordingly. If someone wants my help, unless they're a work friend, I ask them for a project code to charge to. If I depend on another team, I depend on my project manager to talk to their project manager to get us a resource from that team. I'm not doing crazy developer reverse engineering another teams product. I look for opportunities to get my team funding and recognition...

          Is the bank building software in the most efficient manner? Hell no. But the bank is just not the kind of place that's going to hire a team of developers and let them build and support a product. Everything thing is a project and cost center. Some organization improvements have been made in this regard, but I'm talking generally here.

          I'm constantly amazed at the number of developers who never seem to adapt to the bank at all. They come in and just keep working as developers never thinking of projects or funding or anything. They'll try not to 'waste' the banks money or be superheroes. Which is kind of okay if you get recognition for it. But if not, just what are you doing. Look around a little.

          And it's not just developers of course. Managers need to adapt to the bank as well. A good manager would know how to 'manage' these developers to keep everything in line with the banks organization structure, while letting them be developers.

          Everyone has got their role, but you need to keep your head up and recognize the organizational structure for what it is. Then do one of the 3 things at the top. And if your manager is not shielding everything from you, then you need to step up your own game to adapt.

          In this specific case, the employees need to be aware of what the AI watcher is doing.

          Side note, I personally think it should be a law for employees to EASILY know how they are monitored and what metrics are being used. Whether you think being monitored like this is good or bad, it should be clearly stated regardless.

          Then the employees need to coordinate among themselves on their response to it. I recall in one of my first jobs, I was on an R&D team and they contracted our a part of the network stack. The code that came back was just insane. Functions weren't used. Code was repeated. I spent god knows how much time getting everything to work. Could these developers have just been that terrible? Part of me was like friggin Indian developers. Later I found out they were actually being paid for Lines of Code. Literally... lines of code. So of course they're going to write more code. That's the system they were operating in. Kudos to them is how I began to look at it.

          So if you're on an assembly line and you know you're being monitored, talk to your fellow colleagues to make sure everyone operates at a reasonable rate. I started doing that as a developer at the bank. I could be done a task in a day, but I'd take 5. No one actually cared as long as the tasks assigned to me were done. Actually I think the scrum master was happier and their stats looked good and everyone was utilized and

          • @scamper_22 Thanks for your interesting post on three options -- including adapting to an organization's social dynamics which you chose in a particular situation.

            Certainly the #2 adaptation approach you outline works for *lots* of people, as described, say, here:
            "The Gervais Principle, Or The Office According to "The Office""
            https://www.ribbonfarm.com/200... [ribbonfarm.com]
            "The Gervais Principle is this: Sociopaths, in their own best interests, knowingly promote over-performing losers into middle-management, groom under-p

    • dam right I have 6 managers and when I make an error with my TPS reports all 6 tell me about it.

    • Some workers say the results are unpleasant ... just what are these people used to . I kinda remember since my fist job straight outta skool was tupperware ... night shift ... (its what you can get not what you want they say) , that was pretty .... euhm ... intense ? constant control , like, supervisors, if you had 'to go' you need to like ring a bell and hope the guy doing rounds gets to you before it's in your pampers and THEN cameras up yow-ass huh, that must be ... unpleasant gods' damn .. GET A JOB , Y
  • They will use it in the dumbest most intrusive way possible and not want to be subjected to it.

  • You will comply. Resistance is futile. Bur, what happens when a worker ages and slows down? Are they thrown under the bus, paid less, or left to do their best based on their aging output capabilities? Time will tell.
    • Oh, well, Citizen, the answer is simple: any machine that has a part that is wearing out, that is compromising the efficicency of the machine, needs to be replaced with a new part. The old part is then recycled. I'm sure the AI can run for days off the combustion of the old part; nothing is wasted.
    • Why should someone doing less get paid the same as someone doing more?

  • by bobstreo ( 1320787 ) on Saturday February 29, 2020 @12:44AM (#59780986)

    for the workers that just grind and grind well. Everybody works at a different pace, but there are always some people who just crank out the work.

    Also, in the case of a safety incident, what rights do the workers have to the footage for workman's comp claims, or safety lawsuits?

  • Bezos unsurprisingly got one of his biological units from China.

    Wonder if he'd object to equivalent metrics being applied to her, say, oh, exactly one lifetime from now.

  • by SuperKendall ( 25149 ) on Saturday February 29, 2020 @12:46AM (#59780994)

    As it mentioned in the summary, being recorded means it's no longer your word against some manipulative co-worker - you have hard video evidence.

    It also for retail workers, means it's not your word verses a customer - they can see from a recording who was being reasonable and who not.

    It even tempers things like office politics (such as they are at that level) by providing video evidence if something needs to go to court.

    The only people video recording hurts, are the kind of people who don't like evidence around if you get my meaning.

    And yes, I have worked customer facing jobs when I was younger, so have all my siblings.

    • by joe_frisch ( 1366229 ) on Saturday February 29, 2020 @12:53AM (#59781002)

      General recording that lets you examine a situation is different from the rather horrifying (to me) idea of an AI monitoring your activity and reporting on your "efficiency". I can see the latter leading to prohibiting actions that make work a lot more pleasant and only very slightly less efficient.

      Whether its workers or managers, no one should need to have their efficiency completely optimized.

      • I can see the latter leading to prohibiting actions that make work a lot more pleasant and only very slightly less efficient.

        It's true. Humans actually become more efficient in fact when they're comfortable and feel safe, as opposed to being rushed, being made to feel uncomfortable and threatened.
        Humans are not machines and we do not work well when we are treated like machines.

        Also, this: Machines are supposed to serve humans, not the other way around.
        In any given situation ask yourselves this: Are you using the tool, or is the tool using you?

        Also, also this: In my experience, humans really do work best when they are aske

    • by jimtheowl ( 4200185 ) on Saturday February 29, 2020 @01:17AM (#59781042)
      Unless the manipulative boss controls the recording and uses whichever extract he wants a meeting to make a point or humiliate.

      Recordings can also be used by the people that control them, and hurt the people who do not.
    • Meh. In retail, they record you because you're the one most likely to be stealing from the business.

  • That's all this is, more surveillance. Digital slavedrivers. Never mind 'automated killbots' in the military, worry about the Digital Slavedrivers, tirelessly whipping the workers until they drop.
    I have never worked anywhere where having your workstation and you and your work watched every single moment went over well, not with anyone. It's bad enough there are so many gods-be-damned cameras and microphones in every gods-be-damned place these days, and fools buy more of them to put in every room of their h
  • by elcor ( 4519045 ) on Saturday February 29, 2020 @01:54AM (#59781088)
    of using yogic terms to nefarious end
  • by Tablizer ( 95088 )

    Trust me, stay away from the pod bay doors.

  • The supposed smarts aren't even mentioned. It's just simple look back at what happened and work through the issue as a team.

    Maybe it can count products with cameras. Big whop-dee.

  • by bferrell ( 253291 ) on Saturday February 29, 2020 @03:26AM (#59781188) Homepage Journal

    "Efficiency experts" did this is days gone past.... They're called time/motion studies. We didn't like them then and they didn't result in any additional efficiency so the whole concept was s**t canned... Until "do it with a computer" came along to revive all the old tripe.

    • "Efficiency experts" did this is days gone past.... They're called time/motion studies. We didn't like them then and they didn't result in any additional efficiency so the whole concept was s**t canned... Until "do it with a computer" came along to revive all the old tripe.

      No, this is something very new. The "computers" have a secondary role that the "efficiency experts" did not. The computers are learning how to do the job. Machine learning can rely heavily on observed patterns, on training data. Truth be told, such training is probably the primary goal, the worker efficiency the secondary goal.

      From the article: "Many jobs in manufacturing require dexterity and resourcefulness, for example, in ways that robots and software still can't match"

      One of the goals of machine

      • by f00zbll ( 526151 )
        Especially if they are using posenet (pose estimation) and then translating that to robot movements. With improvements to bi-ped robots and boston dynamics wheeled 2 handed robots, those translated movements could be used to teach the robots better ways to fill a shit-ton of orders quickly.
  • by tanveer1979 ( 530624 ) on Saturday February 29, 2020 @03:32AM (#59781192) Homepage Journal

    https://marshallbrain.com/mann... [marshallbrain.com]

    Replace the person managing the workers, instead of replacing the workers.

  • Well, something had to replace slavery. Meet the new massa. Same as the old massa.
  • I think this shows that just like money and guns, the evil is the user not the system.
  • That's how it works.

  • sounds like Taylorism, aka management by numbers, 2.0.

  • The robots have yet to figure out "Beat to shape, file to fit, paint to match."

  • At least data can be subpoenaed vs the word of an overseer.

    Work will suck until workers get angry enough to bargain collectively. The glory days of the labor movement were impressive. Americans had the right (realistic) attitude. Many hated the government which then as today was owned by rich scumbags, and were willing to counter Pinkerton and strikebreaker violence by force. (Anti-2A readers who imagine times have changed forever are summer children.) Labor should be fierce because the elites only respect

    • "Sheltered geeks won't "get" this until things get much worse."

      I agree to a point. There's no need to jump right to smashing car windows and threatening management...that's how unions got a bad reputation and it's the subject of the stories conservatives keep alive. But, there is a need for every worker to realize that labor and management have very different goals. We've spent decades taking down the barriers between these 2 groups, convincing labor that everyone's on the same side. You can only be in one

  • by ErichTheRed ( 39327 ) on Saturday February 29, 2020 @01:58PM (#59782310)

    Employers are rushing headlong into DevOps initiatives because they see their peer companies doing it, they think they can become Google by implementing Kanban boards, or whatever. This exact same kind of 24/7 monitoring has the potential to be misused by crappy micromanagers. I've been doing IT and systems engineering work for over 20 years and the only constant is management's desire to cut costs to the bone, even in places that actually value IT/dev. They'll automate as much as they can (which is good) and even then they'll pit their workers against offshore workers who work for 40% less. I've had upper management come right out and ask me "Why am I paying you when I can get 4 of you for the same price?" I've had good answers to this, but let's just assume that every employer is like this.

    In this kind of environment, the constant pipeline/CICD/deploy/debug/release cycle that gets faster and faster is a goldmine of data for "data-driven" managers in less-enlightened companies. They can easily zero in on a team or an individual team member and bring up detailed statistics on feature velocity, checkin rate, who they collaborate with, etc. All of this is good until they decide to crack the whip harder and guilt people into working longer hours, coding faster, reducing test failures, and so on. A company adopting "unlimited vacation" now has data to beat people over the head with and convince them that they need to keep working instead of taking a break.

    Not every place will be like this. But, I see newer coders who haven't been burned by an employer yet who haven't even thought of this yet. They see an employer providing 3 meals a day, taking care of all their needs as long as they stay "on campus" can can't even imagine that all the monitoring has a potential dark side as well. Wait till the next recession hits and they start using burndown charts sorted by coder to prepare the layoff lists.

"Being against torture ought to be sort of a multipartisan thing." -- Karl Lehenbauer, as amended by Jeff Daiell, a Libertarian

Working...