Will AI take your job?

Spoiler alert: probably not

This question has rocketed through social media algorithms, private text and email exchanges, and quiet collegial conversations in the months since ChatGPT’s first retail release in 2022. It resonates at all levels of all organizations, but most specifically,   people seem to be asking: will AI wipe out the middle manager? 

Middle management is a woebegone role. These managers are simultaneously despised by their reports for climbing the ladder, and dismissed by their seniors for not climbing high enough. They are required to be strategic and emotionally engaged with their work but also implementation-focused and workaday. In tech companies of late, this lowly-yet-not-too-low role has been attacked in the past couple of years. But it is not for the reason of AI. Instead, it seems that tech companies are shedding the personnel for whom they overhired during the global Covid-19 pandemic when, stuck inside and scrolling, much of the world demanded new digital distractions at a breathtaking pace. So much for AI going for middle management first; it seems that upper management still holds sway for now. 

What, then, about entry-level jobs and Ai? While data shows that there is a slowdown in hiring entry-level workers, this does not appear to be due to AI technology as much as it is due to a hesitation of firms to take on new workers while two heavy uncertainties loom: the useful integration of AI technologies into the workplace, and the current lurch-and-shift management of US economic policy. Firms hate uncertainty, and there is nothing more uncertain than a new hire early in their career. Since companies cannot change the uncertainty of AI or of government policy, they logically control hiring.   

This cloudiness will take time to dispel. AI, like all new technologies, needs to be integrated into workflows. Datasets must be redesigned to be AI-consumable; employees must update their thinking and orientation to the advantages of the new tool. To these points, much digital ink has been spilt discussing the decades it took to wipe out elevator operators’ employment, to integrate electricity into factories, and to create the noise and activity of Manchester textile factories. Those arguments state that, although we are not seeing a steep drop off in knowledge work now, it is certainly coming. 

But there is a crucial difference between elevator operators, electricity, and the spinning jenny: those technologies were reliable. AIs trained to be human-like, trained on what humans can produce, will be as reliable as we are, which is to say, not so very much. 

The real danger of AI integration is not to the worker. There is a solid case to be made that, since AIs will remain human-like and unreliable, new middle management jobs will include managing teams of AIs, adding to that already thankless task. Instead, danger to organizations lies in three sectors: 

1. With new, powerful integration tools, new, powerful ways to damage business are already emerging. How much reliance on AIs and thusly exposure to data and other cybersecurity risks, does any business want?

2. AI can speed up productivity relentlessly, but is productivity increasing in the places it matters for businesses? There is solid evidence on social media that the increasing presence of brands in consumer’s faces is already driving them to private chat groups and other, more fragmented versions of social networks. Will clunky, all-over increases in productivity really pay off for firms? After all, a McClaren can hit top speeds of almost 200 mph, but most people simply don’t need that level of power. 

3. With new, powerful, and individual technologies, how much more employee distraction and individualization can firms really afford? AI is a tool employees use by themselves to increase their productivity. Already, memes abound in which one employee uses AI to spin 5 superficial bullet points on a topic into an important-seeming analysis, only to have a recipient employee use AI to break the analysis into 5 superficial bullet points.
   
   This is employee rent-seeking at its worst, but it is also a case of the emperor has no clothes: if every employee, from front-line workers to CEOs, uses AI quietly to make themselves sound more thoughtful and engaged in emails and memos, and everyone also quietly uses AI to strip away that AI-generated nonsense, then no productivity has been gained, no institutional intelligence has been generated. 

Instead, get ready for a human-centered revolution. Those with the ability to speak extemporaneously, charismatically, and succinctly will hold sway in the conference rooms of the future. Listeners might surreptitiously type into their phones replies to oratorical fireworks, but the future of work is performance art of a different and more ancient type than the desk-bound post-war knowledge work to which we’ve become accustomed. 

Instead of winning the argument of the day in memos, emails, and chat threads, get ready for the next generation of business leaders to win by debate and rhetoric. So join a Debate Club, workers of the future, and brush off your old copy of Club rules, middle managers: you might be plunged into AI-team-management hell soon, but it will be tempered by the opportunity to speak persuasively and productively to your human colleagues.  

What we’re into this week

Aaron

I read this article on AI and minimalism back in the summer, and I think its point has only gotten stronger.  AI promises to both create more noise and be a tool against that noise. With opposing forces like that, careers, to your point Ana, are at net zero. I find that kind of depressing, but also, I guess that’s good for us humans.

Ana

For sure, but in the near term, things can be pretty depressing, if only because sometimes orgs are so short-sighted. In the couple of weeks since we recorded this take, there have been quite a few announcements that reinforce what I see happening with AI and careers. Most recently, Meta announced around 600 layoffs from the AI staff. This layoff announcement comes at the same time Meta is hiring for other, related AI jobs as well.

Meta can try to spin this however they want, but reorgs are like rebrands: it's what companies do when they don't know what to do. These layoffs are related to internal politics, not strong strategic vision. The same thing is happening in GSA and across the federal government. 

Scott

That’s true, Ana, and I agree with Aaron, but I’m going to back way up from this conversation. I use AI every day in coding and writing assistants, and I really enjoy having these tools at my disposal. But I can’t say that it feels in any way like my skills are at risk; AI simply makes easier the things that were once far more annoying to me. 

This moment in our history reminds me of Ways of Seeing, this awesome four part series from John Berger and the BBC back in the 1970s. Among many other things, Berger tracks how changes in technology provoke changes in how we perceive both truth and value; I think the same is happening right now in the value perception of human labor. The value won’t disappear, but it will change.   

Ana

I honestly think you’re both right, and that’s backed up by this report from Yale Budget Lab (Oct 1) stating that “broader labor market has not experienced a discernible disruption since ChatGPT’s release 33 months ago, undercutting fears that AI automation is currently eroding the demand for cognitive labor across the economy.” but that more time and data collection is needed to really understand the impact. It’s early days yet. 

Get AI Training

Ishmael Interactive’s AI training arms you and your teams with the foundations of using AIs productively at work. Other AI courses either focus on specific tools or are bland, generalized trainings, but Ishmael Interactive’s training integrates the best of a hands-on workshop with the power of a focused, analytical lecture. 

Our single, 4 hour workshop or four 1 one hour, pre-recorded, a-la-carte trainings with instructor-graded homework, shows your team: 

1. How companies tune AIs and what that means for your use

2. How to use multiple AIs to check against each other

3. How to form the best, structured demands to get the most out of your AI sessions

4. When and how to correct an AI to ensure you continuously see great outputs

We are the only AI training researched and developed for professionals working in mission-driven organizations, by AI researchers who have built and tested machine learning and AI systems for the past decade. Check out How to Think with AI, and get a handle with AI integration at the right level for your business. 

Credits

Guest: Ana Monroe
Interviewer: Aaron Myers
Producer: Ana Monroe
Text: A Adams
Artwork: Plate Number 49. Walking and turning around rapidly with a satchel in one hand, a cane in the other. 1887. Eadweard Muybridge

Why this artwork: Muybridge created hundreds of these photographic motion studies in the late 19th century, demonstrating how the new technology of photography could replace painting and illustration in representational art. While Muybridge was somewhat correct, painting and illustration are both still alive and well. The same kind of upheaval will be felt across professions as AI implementation matures.