Archive.

  • Over 20,000 civil servants were given the latest AI tech for 3 months, using it to draft documents, summarise meetings and more
  • from policy officials using it to cut through jargon and streamline consultations, to Work Coaches speeding up support for job seekers – officials said the tech boosted their ability to deliver the Plan for Change
  • comes as expansive research shows half of office work can be helped by AI, as government continues push to save £45 billion by creating a lean, modern state using tech
  • @Fizz@lemmy.nz
    link
    fedilink
    English
    243 days ago

    26mins seems within the margin of error. You could get at least an extra 26 mins of productivity out of someone in plenty of different ways.

    • funkajunk
      link
      fedilink
      English
      133 days ago

      Ask everyone who the annoying co-worker is; you know the one that yaps to you about some inane bullshit that you don’t care about for up to an hour everyday?

      Fire “that guy” en masse and you would see massive gains.

    • ProOP
      link
      fedilink
      English
      -53 days ago

      You could get at least an extra 26 mins of productivity out of someone in plenty of different ways.

      How so? Can you give an example that is applicable here?

      • @ShittyBeatlesFCPres@lemmy.world
        link
        fedilink
        English
        223 days ago

        Shorten every meeting by 15 minutes and call as few as possible with as few people as possible. Get rid of TPS reports or whatever bullshit forms the English and Welsh government(s) require.

        I mean, fuck, reducing commute times would save people time. Spend the Copilot money on public infrastructure and you’ll increase productivity.

        • @Squizzy@lemmy.world
          link
          fedilink
          English
          13 days ago

          Having non vital people om meetings brings them into the fold, they are not immediately neccessary but its a skills thing. I was dropped strsight into all meetings day one of a new job and it is a great help.

      • @Fizz@lemmy.nz
        link
        fedilink
        English
        113 days ago

        A lot of the tasks ai is doing are process issues. If the worker needs ai to filter a report for to much jargon people need to write more concise reports. There now we’ve saved time for the readers and the writer. For the replying to emails with “personalised” job advice. Thats gross, I dont think its ethical for government workers to give ai responses to people who are desperately seeking employment.

        All kinds of things can motivate workers to care more about their job and thus work harder. Stuff like saying if you finish your work you can go home early (this would net me 4 hours of productivity each day). Things like performance based compensation. Shorter work weeks. Less commute time. Better work environment. Im sure many people could find an extra 26mins in there day if they wanted to.

        Also keep in mind this is a self reported study. It feels productive when ai writes you an email but sometimes you fluff around with ai and you could have just written it 20s.

      • @DoPeopleLookHere@sh.itjust.works
        link
        fedilink
        English
        53 days ago

        Forced overtime comes to an easier, cheaper mind.

        But how better done depends on the field. Me, having a faster computer reduces compile time. So NOT having AI overhead on my machine is more important.

        People could get tons of flows improved by not abusing Excell as a database.

        • Echo Dot
          link
          fedilink
          English
          23 days ago

          People could get tons of flows improved by not abusing Excell as a database.

          We have our knowledge articles in an excel spreadsheet. If you wrote a virus that would scan every single Excel file and just delete any spreadsheet without any formulae in it you would cause enormous amounts of suffering but not do any actual damage.

        • ProOP
          link
          fedilink
          English
          -63 days ago

          I am talking about the study, not in general.

          All your examples show that you did not have anything in mind when you typed your previous comment.

          • @DoPeopleLookHere@sh.itjust.works
            link
            fedilink
            English
            9
            edit-2
            3 days ago

            Your study has no control. So how do we know that’s the best way to get more out of people? The page linked doesn’t even specify what job types.

            I’d still wager they’d be better served by better applications, not AI.

            Also not the same person.

  • BrightCandle
    link
    fedilink
    English
    103 days ago

    the lack of control group undergoing the same monitoring means there is a high chance of bias.

    I would also want to see secondary measures such as the amount of queries people did and what they actually used it for. Might be worth tracking over time to see if there is an increase or decrease in use. These sort of secondary measures give some confidence that it is useful and its continuing to be so and its not just people behave differently when watched.