In contrast to early hopes that AI may become the great leveller in terms of access, the interaction of technology and professional stratification is rarely that straightforward. A recent HEPI report showed that students are sharply divided on whether AI contributes to their learning or hinders it. I find this polarisation not at all surprising, and we are already seeing it with academics.

Knowing how to use AI well, which tool, and for which tasks, is still in the exploration stage. And right now, the debate is dominated by the quants researchers.

With the toxicity of this debate at an all-time high, I agree with Kustov that AI disclosure is simply not viable right now. I heard from a colleague (mostly quants, some mixed methods) that they had something rejected for disclosing their use of AI in a responsible manner. And I do not doubt it. It’s the vibe, and even when a journal has an AI policy, you risk having your work dismissed even if you used AI responsibly and checked everything (which usually takes so long it often negates the purported productivity advantage).

In the BAM AI White Paper, which I co-authored, we called for responsible AI disclosure. But in our recent podcast, I was a lot less confident about that:

So, how well you can use AI will increasingly matter, and it will also make all the difference whether people can tell that you used it or not.

Knowing how to navigate the jagged frontier

One of the best examples of this is Adam Kucharski’s recent blog “How much time did past Adam waste?” — turns out, not that much. And this is for an Excel-based task, where you would think AI tools have an edge:

“What is hard is working with fragmented, incomplete, inconsistent datasets and coming up with clever methodology to tackle genuinely new research questions. And I’m yet to see evidence that common agents are about to takeover in this space. Even if I really do wish they could have saved me all that time a decade ago.”

I similarly have tried various AI tools over the years to see if I could really short-circuit one of the tasks that qual researchers most frequently outsource to the rare research assistant who may come along on a grant or through a co-author:

The structured literature analysis

I have a particular set of articles tightly clustered around an issue, fewer than 50 in number, from which I want to extract very specific information. Originally, I tried Elicit, roundabout 2023. The results were disappointing, and the tool seemed to be set up for medical researchers or similar. Now, with Claude Cowork at the ready, I decided to tackle this again, even though I have already done all the work manually. So it seemed an optimal test along the lines of “how much time could past Stephie have saved?”

Again, not much, as the result returned some egregious errors, which led to a longer conversation with Claude about how it works. In summary, text extraction at volume remains a challenge for context windows, and when the information is not on the first few pages of long academic articles, Claude defaults to “inference” based on training data and known outputs by academic authors.

What I did learn is that my instructions and the memory file for Claude need to include explicit instructions to mark where inference was used and flag for human checking.

If I can get a human research assistant for this task, that would still be superior to Claude. (Sorry, Claude.)

The rare qual voice

Over at Unpublishable Papers, an anthropologist’s view on AI brings up some pretty key issues for any business and management scholar:

“But artificial intelligence has still not changed much about the time and effort that goes into theoretical labor. Theory is still expensive. …But theoretical labor was always the important part.”

Which perhaps sums up why most business and management scholars remain quiet (and smug), even though most research is, in fact, quantitative. But our field is apparently not as empirically driven as the many concerned political scientists and economists (who generally think of themselves as perhaps somewhat superior to the upstart business schools).

The other social sciences, perhaps, sometimes wonder:

“But if theoretical labor is the important part, why don’t we instead have a scientific world maniacally focused on the stakes above all else?

Well, newsflash, everyone, this is what we are maniacally focused on ABOVE ALL ELSE ALL THE TIME.

Not least at the top journals. Sure, there are plenty of low-hanging journals that take the other stuff. They were not particularly influential before. Some significantly expanded volume before the first AI tool was ever launched to make money from the shift to open access. In this ecosystem, the potential for AI to produce work to significantly shift the field remains limited.

Will AI's affordances topple the dominance of theory in business and management? Well, probably not, and for the first time in ages, I think this may be a good thing. If we are lucky, it may even broaden and change the definition of theory, though I am less sure about that.

The only real challenge now is reading smoothly written, well-presented academic prose that even occasionally features a short sentence for impact or distraction, but there is no “there” there.

Of course, being able to write in a specific fashion has long been a shorthand in the field for “quality” (otherwise known as having been trained at the right schools, which are US American or trying to be). But even that, LLMs do not always fully manage. Or, indeed, their users have failed to set up their system correctly because they themselves are unaware of these conventions. Or just very bad at using AI…

Academic Publishing

Now this is where the crunch of the debate is… Whenever I see someone in business and management vociferously complain on LinkedIn (and probably elsewhere that I am not) that they were just sent an AI-generated manuscript for peer review with hallucinated references, I feel:

1. Sorry for the editors, who don’t have the time to check at this level and who are given no support in dealing with an avalanche of AI-generated dross (and who are generally unpaid beyond an honorarium in our field).

2. Angry at the publishers, who make a lot of money, and push their homespun new publishing systems at us with their endless problems, just because they now want to make money from data as well, PLUS licensing OUR writing to AI-companies, and then completely fail to purchase other AI systems that could just scan incoming manuscripts for hallucinated references (yes these firms exist, see for example GroundedAI). OK, breathe…

3. Exasperated with colleagues reviewing these pieces and vocalising their discontent with the moral superiority common to academics. Not because they are annoyed with hallucinated references. But because they then feel they can spot every instance of AI-generated work. You cannot, OK? You can identify the incompetent users at best.

That’s the ground-level experience right now.

Economists, of course, like to take the high ground. They also, it turns out, make their successful paid (ok, probably the archives function), so if you want to read it for yourself, you will need the 7-day trial. I really liked Scott Cunningham’s post, though some aspects really made my eyes roll madly in their sockets.

The first one was this:

Clearly, there is no scarcity in potential output space. Economists may think so, but that reflects distinct disciplinary institutions that reinforce a potentially even more hierarchical journal structure than the one in business and management. Since nobody properly pays editors (though this varies by field), you can hire them just fine. Editors are dirt cheap. The constraints are how many more papers we actually want to publish and whether we can effectively expand quality control.

At a relevant meeting last week, there were some general statements that publishing 1,500 or indeed 2,000 papers per year is clearly not aligned with quality control. It’s a generous assessment; even with a large team, I’d think the boundary for any one journal would be lower. But Cunningham’s argument is about a range of journals, and, as a larger field, business and management has many more of them.

So the idea that there is a limited, not at all dynamic pool of slots being targeted by authors, AI-assisted or otherwise, seems fanciful.

I agreed more heartily with the challenge of human peer review.

So, is human peer review dead? Probably. Is that such a disaster? I totally agree with Scott Cunningham, and so do others like Dave Karpf (now over at Beehiv). Karpf goes further, arguing it is the eulogy of the social science paper. By and large, his argument amounts to Goodhart’s law: we need to consider whether we are measuring the right thing for tenure, promotion, appointment, etc.

“Peer review was already stressed to the breaking point. It DOES NOT SURVIVE when a young researcher can have Claude Code produce a lit review, gather data, conduct a regression analysis, and slap on a passable discussion and conclusion section. Of course we will be flooded by AI-written/researcher-lightly-reviewed articles. Of course peer reviewers will either opt out of the (voluntary, thankless) labor of offering genuine feedback, or will have Claudebot heavily “assist” them in reviewing.

And this is a serious problem for Hiring Committees and Promotion & Tenure Committees. Universities are slow, lumbering bureaucracies. This is an appropriate time for them to freak out and start adjusting to the “death of the journal article.” They’re measuring the wrong thing. They will have to start measuring something else.

That brings me to my second point, though: good riddance!” (Dave Karpf)

What he does not provide is a blueprint for an alternative system. It just turns into another “woe is us in political science, the job market is so bad”. I am sympathetic, but you know? So what?

The economists seem hell-bent on raising submission fees even more as a deterrent, with little consideration for the well-known problems with that approach. Also, what about the rest of us, where submission fees are not accepted? And frankly, if a fee-charging journal would ask me to review for them FOR FREE(!), I would tell them where ot stick it.

So, we are all heading headlong into the academic interregnum:

“The old world is dying, and the new world struggles to be born: now is the time of monsters”. (Antonio Gramsci, Prison Notebooks, c. 1930)

As far as monsters go, I quite like Claude Opus. And you know what I do now when I read something I am fairly sure is mostly AI-generated, with little to no human academic input?

Sorry, this last bit is paid subscribers only ;-) A bit of a personal comment and a “rattle bag” of some great readings out there!

Read more

And in light of how widely (and fiercely) debated this topic became at the start of March, this Friday's post is free to read, because I think this has been quite a lopsided debate.

Lopsided because it has been dominated by the quants, who are (perhaps predictably?) seeing that their methodological skills are more substitutable than they realised.

Also lopsided because the social science debate has been dominated by political scientists and economists. That one had me a little surprised — where are all the other social scientists? Especially business and management?

AI will take all your jobs

The frenzied debate is, of course, easiest to understand as part of the wider debate around whether we humans and our means of survival, otherwise known as jobs/employment, are at stake. As Richard Elsom, aka the AI Archivist, pointed out on LinkedIn, it is mostly people who do not do our jobs and know nothing about them who are busy telling everyone that they will soon be obsolete.

The FT’s AI Shift newsletter regularly digs into claims about AI-induced job-maggedon, and its two authors have currently settled on:

1. Economic data shows that job losses are not due to AI, because you would expect the productivity of the remaining workers to increase — guess what, that is not happening.

   1. There is one caveat to this: Software and Apps have been shipping at a noticeably higher rate since 2025. So the Claude Code effect is real. But by now, you have probably heard about the Jevons Paradox: greater efficiency of a resource leads to greater usage or consumption. So this may not even spell doom to coders. (Also, Google “demand elasticity” and look smart in the next AI debate.)

2. Most papers examining “at-risk” occupations operate on the principle of breaking jobs down into discrete tasks and tracking whether an AI can perform them. But, for the most part, our jobs are not bundles of discrete tasks. Sorry, but simple analytics only get you so far. Also, here is a neat Spider diagram by the FT that will either totally reassure you or give you the willies, depending on what your general inclination is anyway:

   

3. As a historian, you may note many a tortured analogy with past technological revolutions. Electrification and the introduction of spreadsheet software are popular — see the AI as Normal Technology and a previous post here. But the prize for best recent historical analogy certainly goes to David Oks’ elaboration of bank tellers — ATMs — the iPhone. (Worth a read! Click the image below for the full post.)

So, this is a short summary of the general moral panic. But it is so much more entertaining when the argument erupts about your own job.

AI will write your social science papers for you

You may know the post that kicked off the Big Debate. But I want to plug the FT’s AI Shift newsletter again, which discussed how good AI has become at doing quantitative social science analysis. Journalist John Burn-Murdoch reflected that, now that gruntwork can be done by AI, more clever people may be able to run more analysis more quickly, who previously maybe did not have the skills to use R or Python or the time to clean the data. And, of course, then John and Sarah O’Connor discussed the concomitant “AI brain fry”, the exhaustion you get from the speed with which AI solves your problems as your work intensifies.

But, you know, other people…

And then the Substack post that riled most of Bluesky (apparently, how would I know?) appeared:

Kustov IS a social scientist, so this set the cat among the pigeons. And not just with the title and the graphics, but also with a post hoc disclosure that the first post was written by AI… (based on his social media posts).

It’s a good piece, and it has a Part 2 follow-up. Part 1 made one really outrageous claim:

AI can already do social science research better than most professors.

Elsewhere, probably in the Notes which I cannot currently find, he hedged that this refers to the average and points to some of our favourite high-volume/low-standard “academic” publishers. A bit disingenuous, but the reality is that a few years as an editor makes you somewhat disillusioned with academic standards (by which I mean me, though maybe also Kustov).

In part 1, Kustov opines that he will no longer plan for a research assistant in his future workflows. (Ah, our US colleagues… research assistants, what are those?) His excitement for agents also highlights that he is part of a tradition in which workflows can be easily described and outsourced — classically, this is more the case for quantitative researchers.

Overall, Part 2 is perhaps the more thought-provoking piece, which interestingly puts forward a number of points worth discussing:

• “AI exposes what’s already broken in academia and beyond” — honestly, hard to argue with, especially around publishing.

• “Qualitative research and novel data collection will increase in relative value” — let’s all rejoice!

• The “jagged frontier” is real - but it’s also user skill and the tendency of critics to criticise old outmoded models and to be thoroughly unaware of how good the frontier models are.

Especially the last point is noticeable when it comes to the AI doomers in our field, I am afraid to say.

But, what left me somewhat disappointed with the debate was that while there is much debate of what is broken in academia, specifically in academic publishing (a subject I can wax angrily about at all times), and some useful pointers on jaggedness that are real, very little of this really seems to reflect qualitative researchers concerns, or the realities of business and management research.

So let’s look at some of these hotly contested issues from the perspective of a “qual” researcher in business and management.

What are the real bottlenecks in business research?

On reflection, most people would say it is not about producing more papers. Yet many arguments about AI replacing social scientists focus on its ability to produce papers more quickly, including the underlying analysis and the literature review.

Literature review papers are notoriously difficult to publish, by the way, and this will likely get worse now.

In business and management as a field, I would argue, you can publish everything you want, as there are so many journals, so many of them unranked, some questionable, and, of course, you have the predatory ones. And then you can publish anything on SSRN or MPRA, so publishing per se is not a meaningful bottleneck.

But we all know that this is not the game

The game has always been an institutional one: getting into the right journals (the right journals differ by institution type and country, though we all agree more or less on the elite).

This feeds into tenure, probationary review, promotion, or indeed, into avoiding redundancy.

So this means you need a) the cultural knowledge to know the journals, and this is increasingly codified — more on that in a future blog; b) the harder to obtain cultural knowledge and social capital that defines academic communities and what journals (editors & reviewers) want. This type of academic politics may be distasteful when we consider the pervasive gatekeeping behaviour in academia, but let me get back to that in part 2…

What’s your view on the emerging big AI debate? Leave a comment below.

Leave a comment

No, it’s not April Fool’s! I’m releasing one of my popular posts from earlier in the year for all readers — hopefully with some good ideas for you to explore what you can do with the new AI tools!

Did we know we needed a handheld mini-computer that fitted into a pocket and also served as a phone before Apple released the first iPhone in 2008? Not so much. Did you know you wanted to code your own websites and apps? Well, I watched “Whizz Kids” in the 80s (and just so you know, it took many years for a programme to come from the US to non-English-speaking European countries, so I am not THAT old), so… yes, I always did. Not enough to actually learn coding, but with hindsight, that looks like a good choice.

So, after listening to one of my favourite tech podcasts over the snowy weekend, I felt inspired to give vibecoding another go. (They are calling it the Claude Code moment, and I’ve been a bit of a Claude fan from the beginning.) Also, the FT’s AI newsletter, interestingly enough, focused on how AI will change the way social scientists work. I think the sub-text here was quantitative social scientists (it’s the FT, after all, even though they do the occasional bit of qualitative work).

Nevertheless, their point was that it was making working with data easier and quicker, including laborious but thankless tasks such as data cleaning. While there is a fierce debate about the possibilities of AI for qualitative research, and specifically “qualitative coding” (which means identifying themes, not writing computer language here), this type of work has more straightforward quality standards and is very obviously the kind of task a researcher would want to outsource.

The Projects

1. Who has won the Henrietta Larson Prize in BHR?

I thought this would be a straightforward task to check AI capabilities. Well, turns out, it is’t. Only an incomplete list exists, and AI tools differed greatly on how well they rustled up alternative sources from the depths of the internet.

Worth noting that there are currently no records of who won the preceding Newcomen Article Award, as the Henrietta Larson was called prior to 2007. I emailed BHR about that. They don’t know either, but may find out. But even then, I am currently the only one to have won the Henrietta Larson twice. As a paying subscriber, you get the full list (cross-checked and confirmed by BHR).

The tasks that I tested AIs for were data searching and data validation. The tools were Claude and Gemini, and their performance was noticeably different.

2. The personal website, aka the unnecessary vanity project

(Which I really need to update to highlight that I am the only author to win the Henrietta Larson Prize twice. Did I mention this yet?)

My two cents, nobody needs them today with all the profiles and whatnots. Mine is a set of links to my many other profiles, and pulls in my publications through my university’s publication repository (so no additional updating). But it was really super-easy, it looks rather good, and it is completely free to host on GitHub, so what’s not to like?

The task was creating a website (duh) and I only used Claude — it was more a proof-of-concept, really. Also, not everything worked, and I am still not sure why.

3. A global map of business historians: the pièce de résistance

Now, this is a long-running test vehicle of mine, which I roll out periodically to see how useful these AI buddies might be as research assistants. The answer to this has long been disappointing — until now!

Now we have a map. It is not complete. Don’t blame me; relying on publicly available online data sources is trickier than one might expect.

Also, the title is slightly wrong — this map also includes people identifying as management history, historical organization studies and organizational memory. More about how to do this, and what this teaches us about the abilities and limitations of current AI models, after the paywall.

The tasks here were many: data searching, data cleaning, creating an intermediate output, creating a website, and teaching a clueless academic how to work GitHub. I predominantly used Claude with some Gemini and Copilot/ChatGPT thrown in. More on why below.

What is vibecoding?

It’s a fancy word for telling an AI (an LLM to be precise) to write a digital object or programme for you. Sometimes it also writes back with what you need to do or what your options are. That part can be tricky, in my experience, as Google Codespace or GitHub continues to confuse me immensely.

I like to think I am getting better at GitHub due to recent adventures — but better is compared to utterly clueless:

• How do you delete something here? (AI taught me how.)

• What is a branch, and how is it different from a fork? (Still no idea. I could ask AI, but I have not cared sufficiently so far.)

Now, the reasonable question you may ask is: but why?

There are a number of answers:

1. When you ask LLMs for anything, not even that complex, they actually write code to do the thing you asked for. So you are already vibecoding by just putting in queries.

2. If you want a non-human research assistant you can order around (ooh do I?!?) and have this RA do tasks for you that you would find too boring to do yourself, or which do not have high enough priority for you to really engage with.

3. You always imagined a bespoke piece of software that would be brilliant for your research, but which was never feasible to create. Or the AI suggested a brilliant idea you never thought of by extending into your ask. This Substack post is making this point in a much broader way.

For the most part, my engagement with AIs is focused on 2., but I have dabbled with 3. on occasion, though until now found it all not quite as advertised. So I am mostly interested in data searching, data cleaning and data validating.

1. Henrietta Larson Prize Winners

I thought this would be easy. Just a question of data collection online. How wrong I was. Claude identified a multi-year gap, and, by checking various BHR websites and article tags, I confirmed that it was entirely correct. I then went to Gemini, which did an excellent job of finding the missing dates by adding the winners from various announcements by them or their universities, or indeed notes appended to articles in institutional repositories that won the prize.

I then went to Claude to validate Gemini’s list, and Claude picked up on a hallucinated title (my 2018 article was replaced by the title of my 2022 book). Claude and Gemini both struggled with some announcements that confused the announcement year with the year for which the prize was awarded. Gemini also correctly highlighted that the Henrietta Larson Prize was first awarded in 2007, when the existing Harvard Newcomen Prize was renamed. Neither tool could find any record online for previous prize winners.

Finally, I emailed BHR to confirm that the list was correct; that there had been a previous prize; and that research on the physical back copies (or their digitised versions) was required to establish who the winners of the “Newcomen Article Award” were. Apparently, this has been raised before, so watch this space.

In the meantime, here is the only verified and complete list of Henrietta Larson Award winners:

Henrietta Larson Award Winners 2007-2024

What I learned

Gemini is great for anything requiring burrowing into the web. That is not surprising, given that it is a Google product. But in this task, and most others, I find that I regularly hallucinate. Usually once per chat. Usually, by sycophantically attributing papers to me that I never wrote (so very easy to spot most of the time).

Claude is not so great at web searches, but is way better at checking, careful in terms of what can be known, and shows the underlying data. I cannot even recall when Claude last hallucinated - I accused it of a hallucination, it apologised, re-checked and then confirmed it had been right all along — and I had to eat my words… For data validation and research tasks that require high reliability, I would always use Claude (so I mostly use Claude).

2. Personal Website

OK, that was just a bit of fun after I listened to the Hard Fork podcast, where both hosts vibecoded their own websites (which I checked out) and were raving about Claude. I don’t need a personal website, but I wanted to try it out.

Claude does vibecoding amazingly. I see why they call it the Claude Code moment (even though your average joe like me does not use the Code interface). Though I did not use the Claude Code terminal, which I found too intimidating, and just used the normal interface. Hard Fork has a little YouTube tutorial to build websites with the actual Claude Code if you want to use that.

With a few iterations, I had a website. True, the Substack feed panel did not work. But with no problems at all, Claude checked the various places where live updates of my publications live, determined how to pull my publications from my University’s institutional repository, and created a bibliography that is fed from another database and will require no updating. Neat!

What I learned

You can code most things with Code, even if you are completely clueless. GitHub gives you free pages. The era of personalised niche software has arrived, and you can do what you want as long as your subscription credits last and you have a little patience. Amazing.

Next time you need a project website, you don’t need to bother IT, get a software company involved, or pay for WordPress — just get a subscription to an AI tool.

3. The OHN Global Map

Now that was so much fun. Really, this builds on tasks I have been asking AIs to do for the past 8 months to get a sense of how fast their capabilities were improving. There was a lot of hype, and many tools were not nearly as good as the hype.

The most interesting thing for me was that AI tools are not that great at simple internet searches as one might assume. Every couple of months, I have asked various tools to create a directory of business historians in the UK. It sounds straightforward, right? We all have online profiles at our universities, and for some of us this says that we are interested in business or management history or related terms. Web-scraping technologies have been around for a long time, so this sounds like a straightforward ask.

Before you freak out, tools continue to perform really badly at this. Claude fabricated people, Gemini found an odd collection, same for ChatGPT/Copilot, and both tend to place people at universities where they worked many, many years ago (suggesting they mostly rely on ingested training data). At some point months ago, ChatGPT suggested coding a map from the directory. It was terrible.

This time round, Claude instead came up with a plan:

1. Use Harzing’s Publish or Perish software

2. Search Google Profiles

3. Use keywords

4. Export the results

This had so many advantages — it only involved people who defined themselves within the field and made this information publicly available. Google Scholar limits how much information you can request at one time, and the PoP software is designed to manage around it, but has its limits. So don’t expect to get more than a few hundred (low hundreds). Basically, if you did this for a large field, it would not work very well.

After some trial and error, I could export a list to Excel, but it required a lot of data cleaning, like separating first and surnames, etc. I started doing this manually, but with different types of entries and international naming conventions, it was just mind-numbingly boring.

And then I thought — hang on a minute… didn’t the FT’s AI Shift say AI was great for data cleaning? I gave the job to Claude, who fixed it beautifully. The remaining one or two mistakes were easily fixed.

I then had it merge some of the old searches into the document. I asked Claude to validate and check on individuals — that seemed to be trickier for an AI tool. I did some manual checking, but by no means all.

Then I asked Claude to code the map and tell me how to put it online. It explained how to work it on GitHub, and like magic, it actually did. Even though I barely understand how GitHub works.

When I decided to contact a few colleagues to allow me to show their details on the map, it became a little trickier still.

AI tools cannot “see” the websites they create. Which explains why some stuff they do really well, and others that you could give to an undergraduate RA to sort out, they struggle with.

Basically, Claude told me repeatedly that the dots now don’t overlap and work well, even though they really didn’t. I had to insist for several rounds, iterating the map to get something more useful. In the process, I learned more about HTML, CSS styling, and how to update code documents on GitHub.

What I learned

This was great fun and a bit of a hobby. My next step will be to use proper Claude Code and see if I can automate the updating of the map, now that I am crowdsourcing entries and permissions (including the few people who requested removal from the map — though I strongly suspect they are not on there anyway, but I will need to check).

AI boosters claim that these tools are super powerful. The AI doomers sometimes may actually believe that they are more powerful than they really are, and end up either paranoid or disappointed. It is clear that there is software on the web that restricts AI access (robot.txt and other tools, which Claude respects), and that the way that AIs can interact with the digital world is not the same as humans.

AIs read code, humans see the visualisation that the code creates. So AI solutions tend to be more overwrought than human ones, but that is balanced by processing speed.

Does it save time? Maybe, but all of these examples are things that I would never have done manually. This is really more about expanding what I can do, and figuring out where and how that would be useful.

Hopefully, this was useful to you, and do try out some of this stuff. Beats playing video games for entertainment in my view!

In the meantime, though, I need your help.

This map is far from complete, because not everyone actually uses the terms I used for searches (and trust me, those searches were more convoluted than you might think).

For now, I use “business historians” as a shorthand for “historical researchers interested in business, management, organizations and organizational memory”. Don’t be cross, I am writing this in my spare time, and this will be a thousand words longer if I use that label EVERY SINGLE TIME!

Consequently, this map features a fair few anomalies:

• The UK has twice as many business historians as the USA. (Hey, Americans, are you going to let that stand?)

• Argentina has NO business historians. (Guys, don’t you use Google Scholar?!?)

• China has one. (Hardly plausible.)

• Japan has 18. (There should be hundreds…)

I could go on. The data for Europe seems a bit better. That said: Bernardo, you are not currently on the map. Also, can you do something about the Latin American coverage?

So, here’s my ask to you:

• Add yourself to the map (even if you're — maybe — already on it). Right now, the map features only numbers, not names.

• If you want to see yourself on the map, confirm that you are happy for your information to be displayed on the map: your name, your university affiliation, and your research interests (optional).

• Pass this on to your colleagues who also should be on the map.

Add yourself to the map!

The data

Where does the data come from?

All data for this map were compiled from publicly available digital sources. The majority of entries were drawn from Google Scholar, where academics can enter their subject interests. Additional searches were also conducted, but yielded fewer and less reliable results. Some of these results were included.

What data underpins the map?

The Github files only contain numbers and no identifying information, unless released by the person.

Where is the data stored?

The data is stored in a secure folder on an institutional OneDrive, which is GDPR-compliant. If you want to know if your name is included, submit the form. If you want to be removed from the underlying dataset, fill in the form and I will confirm removal via email. All data from the form submission is stored in the same OneDrive folders, and the form is also part of the institutional Microsoft 365 subscription and GDPR-compliant.

What now?

Fill in the form and add yourself to the map. If you feel like it, opt in to display your information (realistically, we are semi-public as professionals already). Share it with your colleagues, so we get a better map.

But why?

Wouldn’t it be nice if we all knew who we are, where we are, and what we do?

It’s not that I disagree. I increasingly appreciate my colleagues posting on LinkedIn (by now the only social media platform I follow professionally, in addition to Substack) when they publish something new, or when a new call for papers is out, or when they summarise their insight or learning from a seminar/conference/workshop. There is so much going on.

Thank you for giving me a short (but not Twitter-short) précis that makes me feel better informed. I genuinely appreciate it.

Research communication for academics

And video abstracts are nice, but I often lack the time to watch a video, even if it is short, and when I have the time, I might be travelling or commuting. But then, I am not of the YouTube generation. Nevertheless, academics talking to camera is not super-exciting. Because of that, I tried my hand at animations for my video abstracts, and that seems to have evergreened them to an extent.

But it takes a lot of time. It was fun to do once or twice, but as a free service, it is too costly in terms of my time.

And then, I thought, what I really want as a consumer of academic literature is a visual abstract that is aesthetically appealing and informationally effective (i.e., reasonably dense). So I have been experimenting with tools like Canva, but tbh I ran into the same problem as with video abstracts. Who has the time?

Once an article is accepted, proofed, online published, and subsequently print published, there is another urgent deadline for revisions, more proofs to correct, and likely conference deadlines (or never-ending internal bureaucratic deadlines for something or other)…

Essentially, once the piece is published, everyone pretty much needs to move on to the next thing. Let’s not forget that, unlike commercial authors, we don’t get paid directly for our writing.

I’ve long thought that what I want is someone to handle my social media, generate some nice visual abstracts, run them by me for correction and approval — basically a kinda research assistant focused on the dissemination aspects of my role.

Enter the AI

Amidst the anxiety of how it may take our jobs, our brains, etc., my first thought was: Ah, finally, the research / personal assistant who can take all the stuff I’m not so keen to waste my time — even better, someone who will never complain about the mind-numbing nature of it.

After the paywall, some tips & tricks to find the right AI to do your research communications for you… And yes, it finally can!

Read more

At some point, I started following the Generative Histories Substack here, which is run by a Canadian historian who does a lot of computer-aided transcription of handwritten documents. During our email archives projects, we had more opportunities to connect with archivists and historians who were using AI tools to transcribe archival sources. Since then, I have loosely followed progress in this area as we used Transkribus to process some files from The National Archives (TNA) in the UK.

The other stream

I also quite enjoy listening to the Hard Fork podcast from the New York Times (free to non-subscribers), which features Casey Newton from Platformer and Kevin Roose from NYT. And the streams crossed (yes, it’s a movie reference, I am that age) — Casey and Kevin were talking to Mark Humphries from Generative Histories about one of his recent post, highlighting that he came across a model that did amazingly well transcribing some tabular data (basically an account book in the old English money with the shillings and pence).

He thought this might be Gemini 3.0. Then Humphries was invited to talk about Gemini 3 on Hard Fork — honestly, that felt just slightly amazing following a historian blogging and then listening to him on a tech podcast. On a fairly arcane commercial issue of a sugar loaf purchase, no less.

They discussed on the podcast that this could be Gemini 3. Then Google released Gemini 3. Then Sunder Pichai went onto the Hard Fork podcast and more or less confirmed that he thinks that the case of the sugar loaf must be Gemini 3. Then Humphries published another blog confirming that Gemini 3 is indeed very good at transcribing handwritten documents. Then, with the inevitability that you might expect, I went to Gemini 3, uploaded a random file from the T70 TNA collection for the Royal Africa Company, and had it transcribe it.

After the paywall, the results of my little experiment…

Read more

Increasingly, the Doomers/Boosters narratives are being replicated by qualitative researchers in business and management. A bit closer to home, it is still entertaining, but also more annoying. And it is becoming increasingly difficult to be heard in the din.

Subscribe now

It’s not surprising, though. The big question is whether AI will be to qualitative research what Excel and spreadsheets have been to quants. That is, will AI tools fundamentally change our practices and make us faster, more efficient and more accurate than before? Because arguably, computers calculating for humans have done that.

The Future, Now and Then

Are large language models on the trajectory of word processing or digital advertising?

In his newsletter this week, Steven Levy looks at the trajectory of ChatGPT and sees echoes of VisiCalc. VisiCalc was the “killer app” of personal computing, the forerunner to a transformative general purpose technology. And Levy has been a tech reporter for four decades. He’s one of the best in the business. He even…

Read more

2 years ago · 60 likes · 11 comments · Dave Karpf

When I still worked at Aston, I had a colleague, an economist by training, who still had hoards of paper files in his office (intimidating even by the standards of a historian), and some went as far back as to the days when you had to hand-calculate your regressions.

(But then, we should not forget that researchers who are not specialists in software programming often use such programmes poorly. So a trade-off of calculative accuracy vs. user error.)

This week, I attended yet another webinar (it’s a bit of a theme at the moment, I realise that), this one about AI and qualitative research, with Stine Grodal and Henri Schildt. Two colleagues we can confidently describe as AI boosters.

Below, which is for the paid subscribers only, I will not only talk about the event but also some of the recent publications by the AI Doomer camp. And a frankly embarrassing report by the UK’s Higher Education Policy Initiative, which is in a category of AI Booster all of its own, and one that I imagine even Stine and Henri would shudder to be associated with.

So, here comes the paywall.

Read more

But more on that later.

I was also watching subscriber numbers with great curiosity as they were getting closer to a round number — 500, here we come! Obviously, that is small for a newsletter on here, but for such a niche topic like organizational history, I think this is not bad. Not bad at all — even if I do say so myself.

And then there was one of those strange moments when the newsletter gets the fleeting “Rising in…” status - it has happened a few times, but honestly, I can never figure out what it means or what it is based on. Nevermind.

A few days later, OHN hit the 500 mark - I was so pleased. And meant to post.

But the very next, they gave it away… 500 was a mirage, never reached, when I looked at my statistics. Not sure what happened there. Really, we are still a few shy of 500 on the free subscribers. Anyone else you can think off to send this to?

Share

Now we also have the little coin/flower/whatever symbol for those of us subscribing to paid. Good to know.

Does Substack make sense?

Much debate on here about whether it does. It’s fair. I subscribed to Wired this year - they offered an annual sub for a rock bottom price, and I find their journalists excellent. Most newsletters here cost way more, and arguably deliver a lot less.

It’s worth highlighting that £5 monthly is the lowest you can charge, equating to £60 annually. To me, that feels steep, so I set my annual price at £20. Many academics don’t charge for theirs. Fair enough.

I have my reasons.

1. The basic newsletter with updates is still free and will always remain free.

2. I do not need to write additional blogs. Frankly, I will only do so if there is a reason. Getting paid is a reason.

3. When I ran OHN on WordPress, after the grant money ran out, I paid for the site rental and the WordPress subscription. Several years. I moved to Substack for the free service (will it remain free?), and I launched “paid” to make some of the money back. Fair and square.

4. I have a problem with Free and Open Access. Yes, it is convenient and drives citations. And every society on the planet right now has a problem with the free internet and the associated clickbait-y enshittification of information and public thought. Free devalues knowledge. It’s time we get real about this. It is not coincidental in a capitalist society. We work in business schools. Tell me one area in business that you would advise anyone to give their core capabilities away for free (unless your game is standard setting)?

Of course, enshittification, originally about platforms, is quickly becoming the Theory Of Everything. International Politics, and UK HE (and life for young people in general). Some debate whether Substack will go for enshittification.

There are a few countervailing forces currently built into the platform: direct access to the content you subscribed to, and creators being able to reach their audience, not mediated by an opaque algorithm; competition — very successful Substackers leave all the time, so competition is alive and well for now.

5. And I have a fifth and final — my last few months as a British Academy of Management Executive & Trustee will be focused on launching a Substack publication for BAM, which will have a multi-author team. So I hope my learning here will help with launching a much bigger, more diverse type of Substack publication. More in due course.

So, onwards now to the paid portion of this message.

Subscribe now

Read more

This week, I am a bit late due to the aforementioned many, many MSc dissertations and other work-related issues. But this seemed a good moment to get back to Notebook LM. Part 1 focused quite narrowly on some of the playful elements and just gave one tight example of an article. But Notebook LM is actually a far better tool when dealing with multiple articles.

Subscribe now

AI ethical considerations

As I said previously, Notebook LM states that what is uploaded is not used for model training, but I generally try to limit my uploading to Open Access articles. That said, most academic articles have been grazed by AI crawlers through different means, which we already know, so the additional damage through an upload here is probably minimal. Academic publishers increasingly license their content to AI companies, like Wiley (though I am unaware that these additional income streams are passed on to academics or scholarly associations owning the journals and providing most of the free labour…)

In several of the notebooks that I will discuss, I actually used the search function within Notebook LM to help me find related articles — so these are available online already.

Notebook LM for literature reviewing

Notebook LM’s features make it a very good tool to help with literature reviewing, in my opinion. Literature reviewing is, of course, an exercise of getting to know the literature, and that is not something that can be easily outsourced.

However, not all aspects of the literature that are discussed in a review need the same level of close engagement. In my other post about literature reviewing, I used the Venn diagram to highlight how to structure them:

The centre and the overlaps are key to carving out one’s argument and contribution. But the wider circles of the different literatures being brought together, few master all of them to the same degree.

• Maybe you are going back to a literature you read a long time ago and do not remember very clearly.

• You may want to take an idea into a new literature.

• Or you are trying to map out how an idea was translated into a different literature (in which you are not so interested) to see what you can learn from how that was done.

This is where Notebook LM is quite useful.

Read more

And, I was impressed. I queried it about academic literature and asked for suggestions. It returned a whole extensive literature on the subject that I had been unaware of, as it has different disciplinary affiliations. Now this is exactly what I want an LLM for!

Subscribe now

This may not be news for those who already had a subscription to the ChatGPT research-level model, but I didn’t. I am not a fan of OpenAI for several reasons, but I specifically dislike that everything I upload will go into model training. Therefore, I tend to use Claude and Notebook LM because these tools indicate that they do not take any uploaded text into their training data. Obviously, I only have these companies’ word for it, so what I upload is my own writing and not that produced by anyone else.

But then we had news at work that MS Copilot will now use GPT5. Copilot at work remains confidential, which means it should be even better than Claude or Notebook LM, but unfortunately, it is not. (Also, our Copilot is not integrated with our Email or OneDrive or Calendar or anything, so it is honestly not that useful.) Yet with GPT5, Copilot is a new proposition.

So I went for a couple of rounds on Copilot/GPT5 these last few days to see if it can now do the things that I want from an LLM:

1. Monitoring new literature on a subject

2. Do the boring webscraping research, which I don’t want to spend time on

3. Generating infographics and presentations from my research papers

Don’t I want it to code my data and also find the stuff that I dimly remember but no longer know in which document/image/whatever it was? I very much want that, but I am concerned about data privacy, and until I don’t feel that this is adequate, that is just on hold.

Monitoring the literature

I asked Copilot/GPT5 whether it can create a literature monitor. I also asked Manus, which brands itself as an AI Agent service. Manus actually generated a page that looks ago and has functionalities. However, they do not deliver anything — it’s a literature monitor that is not really monitoring anything or delivering any outputs: https://19hninc83y8d.manus.space/

It would be nice if it worked. But it doesn’t. And I am not a software engineer — that is the point after all.

Read more

Did you know that AIs loved the Bauhaus? No, me neither…

I assume everyone’s feeds continue to be filled with missives offering the best AI solutions for everything, including research. Normally full of hyperbolic statements, bold type, bullet points — and probably em dashes. (Now that I see more em dashes in my life, I’ll be honest, I find them visually very attractive and for the first time consciously use them…)

So I thought … why not a running commentary of my various experimentations with AI tools and what I find them good for (and not). I won’t bother with spell checkers, but as I remarked elsewhere that RAG Pipelines and Notebook LM definitely have their uses for qualitative researchers involved with lots of text and media data, why not start there?

Notebook LM for researchers

My attempts at vibecoding have been lacklustre so far, so I won’t be building any sophisticated tools here (for those of you with ambitions in this space, AI Archivist is highly recommended!). Notebook LM, however, is free (yes, yes, I know) and the tool claims not to train Google’s models on what you upload (ok, we have to believe a big tech company, but it’s better than OpenAI, which very much gives no such undertaking).

Subscribe now

So, because of the above, I’ll be using one of my recent co-authored articles for this little exploration — and it is open access, which means every LLM out there has already grazed it for free. (And if that does not fill you with joy, here are a few more reasons why you should feel unhappy about that.) It’s also a good idea to use something you know really well, because this will give a quick understanding of both the advantages and limits of the tool.

So use something that you wrote, recently enough that you remember it, which for me is this one:

Decker, Stephanie, Elena Giovannoni, and Emmanuella Plakoyiannaki. (2024) ‘A Microhistory of Architecture: Historical Imagination and the Bauhaus’. Management & Organizational History, advance online: https://www.tandfonline.com/doi/full/10.1080/17449359.2024.2423095.

LLMs as data extraction tools

First up, my perspective on AI, or more specifically LLMs, is that they are a data extraction tool. And as a qualitative researcher, ever time-poor and overloaded with too much to do and to read, I can see the value in expanding my reach through such tools.

Read more

Subscribe now

We could probably already have AI archivists – that does not mean that you won’t have a human with oversight of the technology, but that you won’t be solely dealing with human archivists.

It's also been clear for several years that humans would not be able to deal with the overwhelming amount of digital sources that are likely to be archived in the future, especially if the collections are not clearly structured and do not come through records management processes. In fact, there is a “digital heap” – digital data that has accumulated over the years when there were few to no archival tools available to manage the transfer of digital materials into archives. While tools existed to look for known sensitive data – phone numbers, social security numbers, email and postal addresses – there was no way to identify problematic content and information that could not be made available without having a human read through all of it. Given the volume and sheer disjointedness of some digital collections, this was not going to be feasible.

Read more

In our new paper, we explore how different researchers use born-digital archives. We've identified four distinct user types: aggregators looking for statistical patterns, synthesizers seeking thematic connections, fact finders hunting specific information, and narrators reconstructing historical stories.

The development of AI tools to help these users often happens randomly—like the Garbage Can Model suggests—rather than through careful planning. We argue that we need more opportunities for users, archivists and tech developers to collaborate, so tools can address everyone's needs, not just focus on preservation or on the needs of large-scale data users crunching numbers.

Here is the abstract - you can read the full article open access here.

The benefits of AI technologies in archival preservation are well recognised, though questions remain about their integration into existing processes. AI also shows promise for enhancing user experience and discovery in accessing born-digital materials. However, a limited understanding of the diverse methodological needs surrounding born-digital access risks the creation of one-size-fits-all solutions that suit certain approaches and research questions better than others. This article reviews current efforts in born-digital access and applies the Garbage Can Model from organisation theory to conceptualise the challenge of developing AI-based tools for multiple user types, highlighting the iterative and often decentralised nature of multi-stakeholder decision-making. We address this challenge by creating four born-digital archival user types—the aggregator, the synthesiser, the fact finder, and the narrator—each with distinct motivations and research approaches. Finally, we identify some new opportunities for stakeholders to inform how AI-based tools can be developed to better meet the variety of methodological needs that exist in relation to born-digital archives.

Share Organizational History Network

The article is available open access here: https://link.springer.com/article/10.1007/s00146-021-01369-9 

For more information on AURA and their events follow this link.