"AI Doesn’t Reduce Work—It Intensifies It"
-
"AI Doesn’t Reduce Work—It Intensifies It"
AI Doesn’t Reduce Work—It Intensifies It
One of the promises of AI is that it can reduce workloads so employees can focus more on higher-value and more engaging tasks. But according to new research, AI tools don’t reduce work, they consistently intensify it: In the study, employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so. That may sound like a win, but it’s not quite so simple. These changes can be unsustainable, leading to workload creep, cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems. To correct for this, companies need to adopt an “AI practice,” or a set of norms and standards around AI use that can include intentional pauses, sequencing work, and adding more human grounding.
Harvard Business Review (hbr.org)
The title is an oversimplification (it can reduce some work), but overall the points really resonated with me.
Humans are not machines. We need to pause. We benefit from thoughtful reflection. A deep understanding of our projects that comes from being hands-on is valuable. We have bad days when taking time to work on things that are gentler to our brain is important, even if AI can do it faster.
Otherwise we burn out.
-
"AI Doesn’t Reduce Work—It Intensifies It"
AI Doesn’t Reduce Work—It Intensifies It
One of the promises of AI is that it can reduce workloads so employees can focus more on higher-value and more engaging tasks. But according to new research, AI tools don’t reduce work, they consistently intensify it: In the study, employees worked at a faster pace, took on a broader scope of tasks, and extended work into more hours of the day, often without being asked to do so. That may sound like a win, but it’s not quite so simple. These changes can be unsustainable, leading to workload creep, cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems. To correct for this, companies need to adopt an “AI practice,” or a set of norms and standards around AI use that can include intentional pauses, sequencing work, and adding more human grounding.
Harvard Business Review (hbr.org)
The title is an oversimplification (it can reduce some work), but overall the points really resonated with me.
Humans are not machines. We need to pause. We benefit from thoughtful reflection. A deep understanding of our projects that comes from being hands-on is valuable. We have bad days when taking time to work on things that are gentler to our brain is important, even if AI can do it faster.
Otherwise we burn out.
@pleia2 I suspect trying to work with AI is like trying to get an answer from an automated phone service. There's lots of trying different things and not getting relevant answers and you come out of it exhausted.
(The advantage of a phone tree over an AI work assistant is that you can almost always give up on the phone tree and start pressing combinations of 0 and * and # until it connects you to a human.)
-
R relay@relay.mycrowd.ca shared this topic
-
@pleia2 I suspect trying to work with AI is like trying to get an answer from an automated phone service. There's lots of trying different things and not getting relevant answers and you come out of it exhausted.
(The advantage of a phone tree over an AI work assistant is that you can almost always give up on the phone tree and start pressing combinations of 0 and * and # until it connects you to a human.)
@robotistry I spent a few weeks filling in for an engineering lead while he was on vacation, and reviewing AI-generated code coming from the team was EXHAUSTING.
Developers were being pressured to do more, so they couldn't spend the time getting the deep understanding of the code and solutions, and the onus landed on reviewers to try and figure it out. I just kept throwing it back at the developers to tell them to explain it better. It caused tension and they still had to do the tedious work.
-
@robotistry I spent a few weeks filling in for an engineering lead while he was on vacation, and reviewing AI-generated code coming from the team was EXHAUSTING.
Developers were being pressured to do more, so they couldn't spend the time getting the deep understanding of the code and solutions, and the onus landed on reviewers to try and figure it out. I just kept throwing it back at the developers to tell them to explain it better. It caused tension and they still had to do the tedious work.
@pleia2 @robotistry AI-generated documentation for existing code is the best thing ever! AI-generated pseudocode for well-documented public APIs is pretty good. AI-generated code for new HW described by poorly written proprietary datasheets is not happening in 2026.
AI can't manage even code review for Linux kernel."We need to pause. We benefit from thoughtful reflection. A deep understanding of our projects that comes from being hands-on is valuable." From the point-of-view of _Thinking Fast, Thinking Slow_, I believe that you're saying we tend to process AI-generated code with System 1, but to make real progress, we need to spend quality time with System 2.
-
R relay@relay.an.exchange shared this topic
-
@robotistry I spent a few weeks filling in for an engineering lead while he was on vacation, and reviewing AI-generated code coming from the team was EXHAUSTING.
Developers were being pressured to do more, so they couldn't spend the time getting the deep understanding of the code and solutions, and the onus landed on reviewers to try and figure it out. I just kept throwing it back at the developers to tell them to explain it better. It caused tension and they still had to do the tedious work.
@pleia2 I was thinking more along the lines of clerical staff being told they have to add AI to their workflows and trying to get it to help populate spreadsheets and turn them into reports. Reviewing AI-generated code sounds awful!
