First, congratulations on a strong article. This captures a lot of what has happened to the secretary pretty accurately. It understates how much her social management tasks have been split off to administrative and personal assistants, who believe me are still dogsbodies in the fine old tradition of "honey get me a cup of coffee" and "don't tell the boss I don't have his report ready." It underrates the role of the word processing center operators, who convert, yes, Word files pre-typed by engineers, accountants, consultants, and executives into a) English b) company format standard c) non-gibberish in a larger sense. But there's a lot of good here.
What's missing?
A recent online article reported on the work of a Marie-Kondo-like efficiency expert who is studying the effect of AI on companies trying to adopt it. She found that in many cases AI was "an HR checkbox to be ticked off" - you attend the seminar, just as you attend a safety or non-harassment seminar, and you go back to work. In others, the AI as purchased is good for stuff, just not our stuff. These are the consequences of executives choosing to buy into AI while having no idea what their companies do and indeed no intention of ever using it themselves.
Her most telling finding, and one that I think is overlooked in this article, is that for any company that has been around long enough to develop employee turnover, how things should be done and how they are done are widely separated.
Dave is the Sr VP on that, and his assistant Madge used to handle that, but when she got overworked Madge admitted she wasn't very good at that so she handed that over to Susan, who traded that task for something she sucked at. Susan left, so that was handed off to Susan's successor Jack, who wasn't good at that but still had to do that, because Madge's boss outranked his. Or Madge left and his successor Bob didn't even know he was supposed to handle that, let alone who in this labyrinth has been doing that all along, so Jack (in his copious free time) teaches the new guy how to prepare the necessary bits of that to hand back to Jack to be handled. Then it turns out Bob and Jack don't like each other, or, tellingly, their bosses don't like each other, and since neither of their bosses does much with his day besides curry his own boss and diligently block communication to his boss's rival's area, Bob is forbidden to ask Jack for help, and Jack is only too willing to let him drown.
AI crashes in, creating rational org charts of tasks and their obvious workflows, which resemble the current company operations not at all. Do we stop everything while AI moves us around like chess pieces? These pesky human relationships do not go away because efficiency.
Second. As someone who used to take minutes and who experienced the transition to chats with verbatim capture, I can tell you that much is said in meetings that ought never, ever to be committed to permanent record. These records are admissible in court, as are texts and every other imaginable form of communication captured forever by electronics, and businesses are learning this to their cost. By giving up their human stenographer, businesses gave up that cushion against liability. At least you could threaten or bribe a stenographer.
Third, you state, "(Although it was still done mostly by women. Economic change does not always imply social change.)" Gender in the workplace is not a social factor, it is an economic factor, as is race, and for exactly the same reasons. See Isabel Wilkerson's CASTE.
Fourth. Managers who monitor their underlings' productivity are redundant and they know it. Like a lot of white collar workers, they owe their employment to a combination of nepotism and a superficial inoffensiveness in the office environment. If they had to catch their own food, they'd starve. Their bosses soon figure out they can be spared.
Fifth. Computerization has not improved standards; it has merely homogenized them. When humans do work, even soul-killing work, they either get bored and get out or they start to slack or sabotage or, in the overwhelming majority of cases, they start to pay attention and make it matter, they get fussy, they figure out how to do it better. When computerization was introduced in the offices in the 80s (I was there) there was more hue and cry among the clerks and secretaries that they were being asked to do a worse job only faster, than among those who objected to learning the computer, and this applied not just to document production / handling and records management but to communication protocols. When companies ordered their clerical workers to fit their duodecahedronal tasks into square computerized holes, data was lost forever, as well as these workers' hard-won, thoughtfully developed methods of tracking and processing data.
Sixth. The office wife was a loss that male workers realized immediately. The loss they didn't feel ... immediately ... was the office mom. Call her Sylvia, a woman I met in the mid-eighties when I walked into a temp gig in a mid-size company. (My friend, you cannot hide company secrets from temps. In a week I knew who was having sex with whom, and in ten days I knew which vendors were getting paid 60-90-never, which VPs were scheduled for the chop, and whether the company would be sold or just closed.) Sylvia taught the new secretaries company standards for their work and the corporate culture. What's more, she taught them to put up with the culture. Where the office mom begrudged or shirked that aspect of her job, you had more sexual harassment cases. In brief, Sylvia taught the girls how to be office wives. I was ordered by the low-level HR manager to "teach Sylvia the computer in three weeks or else she's fired." I told him baldly that it wasn't my job to fire his people, but I would teach her the computer. I also related this full conversation to Sylvia. Next time he came boiling out of his office in a panic, demanding coffee, a cab, and that she block all calls from his boss, she looked up and said, "I'm sorry, sir, I'm using The Computer."
THAT was the revolution. Corporate offices have never recovered from this destabilization.
I could rant further but really, this reply is overlong.
It was the perfect length and you are so right - I missed so much. If I'd gone into it in as much detail as I wanted I would have written a book! I think my mum would agree with you about much of this. I also got a lot from the book Computer Chips and Paper Clips which contains records of 1970s and 1980s managers buying computers in exactly the way you suggest executives are buying AI right now
<laughing> yes, and they did. They bought the computer system whose salesman gave them the best lunch, not the one that would work best for their company, because *they weren't going to use it anyway.*
The irony is that the best use of AI outside of medical diagnosis is at the CEO level of decision-making. And because chief executives often make decisions with the little head, they figure it doesn't apply to them. Even if they're smart, they're unlikely to replace themselves with AI.
What a delightful deep dive and reminder that while this change is happening so quickly, it is not wholly unprecedented. We are all managers now, and love the prediction of what that means for the future of work.
Really enjoyed this piece. One thing the comments have surfaced particularly well is the gap between the formal workflow and the one that actually keeps the place running. Secretarial and administrative labour seems to have carried a huge amount of tacit organisational knowledge: unofficial handovers, quiet repairs, personal loyalties, knowing who really does what and who must never be copied on an email. None of that ever appeared on an org chart, yet it was often what made the org chart workable.
It made me wonder whether the historical analogue for AI isn’t simply task automation, but an attempt to formalise and absorb that tacit knowledge. If that’s right, AI may end up less a labour-saving tool than a very expensive device for revealing how irrational and improvisational most organisations already are.
The gender dimension seems important here, too. A lot of the work didn’t vanish. It was renamed, professionalised, and in some cases revalued once detached from the figure of “the secretary”. I’m curious whether AI will produce another version of that move, where some low-status coordination work becomes “orchestration” or “AI management”, while the people still doing the underlying organisational glue remain largely invisible.
And I do wonder how open-plan offices fit into this history. Secretaries once acted as physical interfaces and filters. The open plan flattened that space and removed a lot of those buffers, which probably explains the modern invention of Teams messages that read like apologies for existing. AI seems to promise a new interface again. The question is whether it restores coordination or simply adds another layer of systems for people to manage.
I was a word processor at UC Berkeley in the 1980s. I had to learn troff and eqn to prepare papers for journals. (This was before LaTEX was popular.) This new upheaval is nothing new, just a different set of programs to learn. I will admit, touch typing stood me in good stead when I became a Software Engineer.
The Harrison Ford/Melanie Griffith/Sigourney Weaver movie, "Working Girl" inadvertently became a capture of that world about to end.
I remember the offices with the architecture shown in the movie. The males with the offices that had windows and views. Just outside their door, a narrow corridor for accessing the offices on one side, and the carels with their female personal secretary for each, on the other.
The movie was made in 1988 - the IBM PC had been out for 8 years. But when I started in a firm of some 4000 desk workers in the fall of 1986, there were only 100 PCs. By 1988, double or triple that, then faster after 1990, but really took off when they could network. I think we had that crucial "computer on every desk" so that nobody had only paper, by about 1996.
All good points, In the 2020s era the admin role still exists but perhaps perception of secretaries for all but the highest c suite have been coded away from simply maintaining office kitchen, coffee, scheduling travel and that sort of thing. I freely speculate that secretary jobs are perceived as being the maid to do that stuff which is perceived as demeaning. As a result that work has been slyly transferred to general staff. Engineers make coffee and schedule all their admin stuff but their billable salary is still 150-300 plus an hour lol! (not what they are paid but their total cost to the company including taxes and benefits.)AI certainly change this.
This is a fascinating article! I learned shorthand and typing in high school and worked after school and during college summers as a secretary for five years. Then I had a secretary for a number of years in the corporate world and, yes, having one was really nice.
It's interesting to compare how AI is already changing day-to-day work with how secretaries became less and less relevant. Although administrative assistants are doing much of the work (minus the typing) of old school secretaries, so in some ways the title simply changed.
The demise of secretaries started long before personal computers had a substantial presence in the office. In the early 1980's, I was Product Manager for Digital Equipment Corporation's (DEC) ALL-IN-1 Office system which ran on VAX minicomputers. At one point, we had almost 50% of the US "office automation" market -- before personal computers had significant market penetration. I can assure you that the demise of secretarial work was not a mere side-effect of bringing computers into the office and teaching men to type. Converting secretaries to administrative assistants was, in fact, a very intentional part of our strategy.
The work that secretaries did was often mechanical and dehumanizing. A woman would graduate from college and find that the only job available to her was mostly mindless copying, typing, taking dictation, etc. Companies like Wang and IBM worked hard to build efficiencies through the creation of typing pools, remote dictation systems, etc. We, on the other hand, strove to automate as much as we could of the office's mechanical non-thinking work in order to move women to the much more enriched and fulfilling job of administrative assistant. Rather than being mere machines, like most secretaries, who had no real career path -- other than "marry the boss," an administrative assistant was an entry level manager who could dream of rising high in the organization. Many did! I think we did a good thing. I know we were trying to. If AI can do for a class of people what "office automation" did for women, it will be a thing to celebrate.
Too many people today look at AI technology as something that will allow them to cut costs and staff -- to do what they do today but more cheaply. I think those people will, in the end fail.
Imagine two CEOs. CEO #1 uses AI to cut costs and staff. But, CEO #2 sees that with AI his existing staff can do a great deal more work, at higher quality, and they can innovate more rapidly. In time, CEO #2's company will grow and hire the workers laid off by CEO #1. Eventually, the only one without a job will be CEO #1.
I agree AI won't impact construction laborers or welders/solderers. But I don't see why their skill levels will go down either. Which illustrates the issue with so many AI projections; they are way too superficial.
Every day, I look around at the jobs I see and ask a simple question; "What will AI replace and how?" I just don't see the massive disruption, and certainly not in 18 - 24 months. Will AI replace the traffic cop, the bricklayer or sheetrock installer, the convenience store clerk, the shelf stockist in the local supermarket, or the worker in the nail salon? It's hard to imagine.
10 years ago robots were supposedly going to transform work but, for the most part, they haven't. So, will AI be another oversold disruption that is met with adaptation, not dislocation?
While the secretary may have reduced in numbers, administrative assistants and executive assistants have taken their place. When working for a lobbyist firm pre-2008, you'd see this transition take place sort of in real time. Secretaries commanded elevated salaries, an executive assistant less, and administrative assistant even less. When things started getting choppy at the end of 2006 and even choppier at the beginning of 2007 due to the Chinese market meltdown and it's spillover to the US markets, the discussions around cost savings per operations came into play. After layoffs, the need for assistants were still there, however the preference to save money was paramount. After 2008, the hiring practice shifted from secretaries to assistants as tech allowed for assistants to do what secretaries did. Another upside too, secretaries were by an large on the ends of their career, they were older and retiring out. Assistants were younger and were to be around longer.
With automation and AI these days there sort of this naive stance that assistants and paralegals will face extinction. However a lot of the people making these claims simply have not worked in the office nor legal offices. To put it simply, the legal field has undergone significant efficiency gains since post 2008. Legal clerks themselves still exist to do the work that the legal assistants simply do not have the time to do. As the legal assistants pick up the work that the paralegals have no time for. The lawyer is no way going to spend his extra time with ai prompts if that takes away from his other tasks. Same goes with office management. They may enjoy the productivity that AI gives, but they themselves do not want to take on the extra burden of prompting AI after firing the assistants. Instead you get less clerks and less lower tier office assistants, but as work increases those positions re-find their place in the office settings.
Perhaps one of the most interesting things I've seen of late is the wage increases of senior executive assistants that are now on par with secretaries decades past.
Welp, good post. It got a rare comment from me, so that should say something. - Best.
I'm honoured! Very interesting thank you. The drop in that chart of admin jobs after 2008 is incredibly steep. Great links. I'll look forward to watching and reading
Didn't mean to come off as arguing against the chart. I suspect the low point where it's at is probably by and large where it'll remain. A lot of the assistant jobs consolidated so the assistant is shared over that of every mid level having their own assistant. Think AI might actually help those in the field over eliminating the final few.
You didn't at all. I was really just saying that the economic data matched your anecdata. Thank you so much for that BBC film by the way, I love it. The robot letter carrier! Perfect
So, so many comments rising to mind.
First, congratulations on a strong article. This captures a lot of what has happened to the secretary pretty accurately. It understates how much her social management tasks have been split off to administrative and personal assistants, who believe me are still dogsbodies in the fine old tradition of "honey get me a cup of coffee" and "don't tell the boss I don't have his report ready." It underrates the role of the word processing center operators, who convert, yes, Word files pre-typed by engineers, accountants, consultants, and executives into a) English b) company format standard c) non-gibberish in a larger sense. But there's a lot of good here.
What's missing?
A recent online article reported on the work of a Marie-Kondo-like efficiency expert who is studying the effect of AI on companies trying to adopt it. She found that in many cases AI was "an HR checkbox to be ticked off" - you attend the seminar, just as you attend a safety or non-harassment seminar, and you go back to work. In others, the AI as purchased is good for stuff, just not our stuff. These are the consequences of executives choosing to buy into AI while having no idea what their companies do and indeed no intention of ever using it themselves.
Her most telling finding, and one that I think is overlooked in this article, is that for any company that has been around long enough to develop employee turnover, how things should be done and how they are done are widely separated.
Dave is the Sr VP on that, and his assistant Madge used to handle that, but when she got overworked Madge admitted she wasn't very good at that so she handed that over to Susan, who traded that task for something she sucked at. Susan left, so that was handed off to Susan's successor Jack, who wasn't good at that but still had to do that, because Madge's boss outranked his. Or Madge left and his successor Bob didn't even know he was supposed to handle that, let alone who in this labyrinth has been doing that all along, so Jack (in his copious free time) teaches the new guy how to prepare the necessary bits of that to hand back to Jack to be handled. Then it turns out Bob and Jack don't like each other, or, tellingly, their bosses don't like each other, and since neither of their bosses does much with his day besides curry his own boss and diligently block communication to his boss's rival's area, Bob is forbidden to ask Jack for help, and Jack is only too willing to let him drown.
AI crashes in, creating rational org charts of tasks and their obvious workflows, which resemble the current company operations not at all. Do we stop everything while AI moves us around like chess pieces? These pesky human relationships do not go away because efficiency.
Second. As someone who used to take minutes and who experienced the transition to chats with verbatim capture, I can tell you that much is said in meetings that ought never, ever to be committed to permanent record. These records are admissible in court, as are texts and every other imaginable form of communication captured forever by electronics, and businesses are learning this to their cost. By giving up their human stenographer, businesses gave up that cushion against liability. At least you could threaten or bribe a stenographer.
Third, you state, "(Although it was still done mostly by women. Economic change does not always imply social change.)" Gender in the workplace is not a social factor, it is an economic factor, as is race, and for exactly the same reasons. See Isabel Wilkerson's CASTE.
Fourth. Managers who monitor their underlings' productivity are redundant and they know it. Like a lot of white collar workers, they owe their employment to a combination of nepotism and a superficial inoffensiveness in the office environment. If they had to catch their own food, they'd starve. Their bosses soon figure out they can be spared.
Fifth. Computerization has not improved standards; it has merely homogenized them. When humans do work, even soul-killing work, they either get bored and get out or they start to slack or sabotage or, in the overwhelming majority of cases, they start to pay attention and make it matter, they get fussy, they figure out how to do it better. When computerization was introduced in the offices in the 80s (I was there) there was more hue and cry among the clerks and secretaries that they were being asked to do a worse job only faster, than among those who objected to learning the computer, and this applied not just to document production / handling and records management but to communication protocols. When companies ordered their clerical workers to fit their duodecahedronal tasks into square computerized holes, data was lost forever, as well as these workers' hard-won, thoughtfully developed methods of tracking and processing data.
Sixth. The office wife was a loss that male workers realized immediately. The loss they didn't feel ... immediately ... was the office mom. Call her Sylvia, a woman I met in the mid-eighties when I walked into a temp gig in a mid-size company. (My friend, you cannot hide company secrets from temps. In a week I knew who was having sex with whom, and in ten days I knew which vendors were getting paid 60-90-never, which VPs were scheduled for the chop, and whether the company would be sold or just closed.) Sylvia taught the new secretaries company standards for their work and the corporate culture. What's more, she taught them to put up with the culture. Where the office mom begrudged or shirked that aspect of her job, you had more sexual harassment cases. In brief, Sylvia taught the girls how to be office wives. I was ordered by the low-level HR manager to "teach Sylvia the computer in three weeks or else she's fired." I told him baldly that it wasn't my job to fire his people, but I would teach her the computer. I also related this full conversation to Sylvia. Next time he came boiling out of his office in a panic, demanding coffee, a cab, and that she block all calls from his boss, she looked up and said, "I'm sorry, sir, I'm using The Computer."
THAT was the revolution. Corporate offices have never recovered from this destabilization.
I could rant further but really, this reply is overlong.
It was the perfect length and you are so right - I missed so much. If I'd gone into it in as much detail as I wanted I would have written a book! I think my mum would agree with you about much of this. I also got a lot from the book Computer Chips and Paper Clips which contains records of 1970s and 1980s managers buying computers in exactly the way you suggest executives are buying AI right now
<laughing> yes, and they did. They bought the computer system whose salesman gave them the best lunch, not the one that would work best for their company, because *they weren't going to use it anyway.*
The irony is that the best use of AI outside of medical diagnosis is at the CEO level of decision-making. And because chief executives often make decisions with the little head, they figure it doesn't apply to them. Even if they're smart, they're unlikely to replace themselves with AI.
Maybe you should write a book.
What a delightful deep dive and reminder that while this change is happening so quickly, it is not wholly unprecedented. We are all managers now, and love the prediction of what that means for the future of work.
Thank you!
Really enjoyed this piece. One thing the comments have surfaced particularly well is the gap between the formal workflow and the one that actually keeps the place running. Secretarial and administrative labour seems to have carried a huge amount of tacit organisational knowledge: unofficial handovers, quiet repairs, personal loyalties, knowing who really does what and who must never be copied on an email. None of that ever appeared on an org chart, yet it was often what made the org chart workable.
It made me wonder whether the historical analogue for AI isn’t simply task automation, but an attempt to formalise and absorb that tacit knowledge. If that’s right, AI may end up less a labour-saving tool than a very expensive device for revealing how irrational and improvisational most organisations already are.
The gender dimension seems important here, too. A lot of the work didn’t vanish. It was renamed, professionalised, and in some cases revalued once detached from the figure of “the secretary”. I’m curious whether AI will produce another version of that move, where some low-status coordination work becomes “orchestration” or “AI management”, while the people still doing the underlying organisational glue remain largely invisible.
And I do wonder how open-plan offices fit into this history. Secretaries once acted as physical interfaces and filters. The open plan flattened that space and removed a lot of those buffers, which probably explains the modern invention of Teams messages that read like apologies for existing. AI seems to promise a new interface again. The question is whether it restores coordination or simply adds another layer of systems for people to manage.
I was a word processor at UC Berkeley in the 1980s. I had to learn troff and eqn to prepare papers for journals. (This was before LaTEX was popular.) This new upheaval is nothing new, just a different set of programs to learn. I will admit, touch typing stood me in good stead when I became a Software Engineer.
This is great
The Harrison Ford/Melanie Griffith/Sigourney Weaver movie, "Working Girl" inadvertently became a capture of that world about to end.
I remember the offices with the architecture shown in the movie. The males with the offices that had windows and views. Just outside their door, a narrow corridor for accessing the offices on one side, and the carels with their female personal secretary for each, on the other.
The movie was made in 1988 - the IBM PC had been out for 8 years. But when I started in a firm of some 4000 desk workers in the fall of 1986, there were only 100 PCs. By 1988, double or triple that, then faster after 1990, but really took off when they could network. I think we had that crucial "computer on every desk" so that nobody had only paper, by about 1996.
All good points, In the 2020s era the admin role still exists but perhaps perception of secretaries for all but the highest c suite have been coded away from simply maintaining office kitchen, coffee, scheduling travel and that sort of thing. I freely speculate that secretary jobs are perceived as being the maid to do that stuff which is perceived as demeaning. As a result that work has been slyly transferred to general staff. Engineers make coffee and schedule all their admin stuff but their billable salary is still 150-300 plus an hour lol! (not what they are paid but their total cost to the company including taxes and benefits.)AI certainly change this.
Thanks for writing this.
I have been rewatching Mad Men again so this piece on secretaries feels timely.
Hard to believe how much the world changed labor wise and how much it will continue to change.
This is a fascinating article! I learned shorthand and typing in high school and worked after school and during college summers as a secretary for five years. Then I had a secretary for a number of years in the corporate world and, yes, having one was really nice.
It's interesting to compare how AI is already changing day-to-day work with how secretaries became less and less relevant. Although administrative assistants are doing much of the work (minus the typing) of old school secretaries, so in some ways the title simply changed.
If you haven't read Dr Carlota Perez's book you might find her thesis interesting
https://en.wikipedia.org/wiki/Technological_Revolutions_and_Financial_Capital
Tl;dr it takes at least 20yrs to work out how to use the technology properly. The managerial and process changes are as critical as the technology.
She’s fantastic isn’t she
The demise of secretaries started long before personal computers had a substantial presence in the office. In the early 1980's, I was Product Manager for Digital Equipment Corporation's (DEC) ALL-IN-1 Office system which ran on VAX minicomputers. At one point, we had almost 50% of the US "office automation" market -- before personal computers had significant market penetration. I can assure you that the demise of secretarial work was not a mere side-effect of bringing computers into the office and teaching men to type. Converting secretaries to administrative assistants was, in fact, a very intentional part of our strategy.
The work that secretaries did was often mechanical and dehumanizing. A woman would graduate from college and find that the only job available to her was mostly mindless copying, typing, taking dictation, etc. Companies like Wang and IBM worked hard to build efficiencies through the creation of typing pools, remote dictation systems, etc. We, on the other hand, strove to automate as much as we could of the office's mechanical non-thinking work in order to move women to the much more enriched and fulfilling job of administrative assistant. Rather than being mere machines, like most secretaries, who had no real career path -- other than "marry the boss," an administrative assistant was an entry level manager who could dream of rising high in the organization. Many did! I think we did a good thing. I know we were trying to. If AI can do for a class of people what "office automation" did for women, it will be a thing to celebrate.
Too many people today look at AI technology as something that will allow them to cut costs and staff -- to do what they do today but more cheaply. I think those people will, in the end fail.
Imagine two CEOs. CEO #1 uses AI to cut costs and staff. But, CEO #2 sees that with AI his existing staff can do a great deal more work, at higher quality, and they can innovate more rapidly. In time, CEO #2's company will grow and hire the workers laid off by CEO #1. Eventually, the only one without a job will be CEO #1.
I agree AI won't impact construction laborers or welders/solderers. But I don't see why their skill levels will go down either. Which illustrates the issue with so many AI projections; they are way too superficial.
Every day, I look around at the jobs I see and ask a simple question; "What will AI replace and how?" I just don't see the massive disruption, and certainly not in 18 - 24 months. Will AI replace the traffic cop, the bricklayer or sheetrock installer, the convenience store clerk, the shelf stockist in the local supermarket, or the worker in the nail salon? It's hard to imagine.
10 years ago robots were supposedly going to transform work but, for the most part, they haven't. So, will AI be another oversold disruption that is met with adaptation, not dislocation?
Your mum was a cutie!
Still is!
Awesome academes report. BBC did a special in 1980 asking a similar question; https://www.youtube.com/watch?v=jRTv9S8ufBw
Kitces had an interesting post the other week making the case for administrative assistants.
https://www.kitces.com/blog/fractional-admin-assistant-hiring-capacity-advisor-efficiency/
I think to the WSJ piece about the 3k an hour lawyer, https://www.wsj.com/business/lawyer-hourly-rate-bill-3400-807cf6ce
While the secretary may have reduced in numbers, administrative assistants and executive assistants have taken their place. When working for a lobbyist firm pre-2008, you'd see this transition take place sort of in real time. Secretaries commanded elevated salaries, an executive assistant less, and administrative assistant even less. When things started getting choppy at the end of 2006 and even choppier at the beginning of 2007 due to the Chinese market meltdown and it's spillover to the US markets, the discussions around cost savings per operations came into play. After layoffs, the need for assistants were still there, however the preference to save money was paramount. After 2008, the hiring practice shifted from secretaries to assistants as tech allowed for assistants to do what secretaries did. Another upside too, secretaries were by an large on the ends of their career, they were older and retiring out. Assistants were younger and were to be around longer.
With automation and AI these days there sort of this naive stance that assistants and paralegals will face extinction. However a lot of the people making these claims simply have not worked in the office nor legal offices. To put it simply, the legal field has undergone significant efficiency gains since post 2008. Legal clerks themselves still exist to do the work that the legal assistants simply do not have the time to do. As the legal assistants pick up the work that the paralegals have no time for. The lawyer is no way going to spend his extra time with ai prompts if that takes away from his other tasks. Same goes with office management. They may enjoy the productivity that AI gives, but they themselves do not want to take on the extra burden of prompting AI after firing the assistants. Instead you get less clerks and less lower tier office assistants, but as work increases those positions re-find their place in the office settings.
Perhaps one of the most interesting things I've seen of late is the wage increases of senior executive assistants that are now on par with secretaries decades past.
Welp, good post. It got a rare comment from me, so that should say something. - Best.
I'm honoured! Very interesting thank you. The drop in that chart of admin jobs after 2008 is incredibly steep. Great links. I'll look forward to watching and reading
Didn't mean to come off as arguing against the chart. I suspect the low point where it's at is probably by and large where it'll remain. A lot of the assistant jobs consolidated so the assistant is shared over that of every mid level having their own assistant. Think AI might actually help those in the field over eliminating the final few.
You didn't at all. I was really just saying that the economic data matched your anecdata. Thank you so much for that BBC film by the way, I love it. The robot letter carrier! Perfect
I have been seeing a new genre of posts and youtube videos, by people using AI. They go roughly as follows.
"Guys! Guys! If you tell AI exactly what you want and in what format, it gives you that! Who knew?"
Provide clear, explicit instructions to get what you expected, how you expected it. This is apparently a revelation worth sharing.
AI will force managers to think about what they want. If AI has no other effect, the productivity boost from nearly adequate management will be large.
Totally although you might have a more optimistic view of our ability to learn management than I do!