Written by Fola Yahaya
Post-Covid, most people are experiencing the negative impact of too much money chasing too many goods. Inflation has engendered a cost of living crisis and I now find myself questioning why in London it costs close to £5 ($7) for coffee beans and water masquerading as a frothy cappuccino. So, as an economist I’m always perplexed by the gap between harsh economic reality and the overly rosy picture painted by the media.
I’ve been calling a global recession for some time, but governments keep shifting the goalposts and eking out a statistical get-out-of-jail card every time we’re close. If you ignore the headlines, and look at what people are experiencing day-to-day, it doesn’t look good. So what has this got to do with AI? Well, a lot. New data from LinkedIn, which tracks how often its users get new jobs, shows a significant slump in white-collar jobs since 2018, and many of these are in sectors that are eminently AI-able.
The decline in certain sectors is unsurprising. HR jobs, for example, have slumped by 28% since 2018. Hiring in marketing, another department that’s often the first to lose its budget in leaner times, is down 23%. But what stands out the most for me is the drop in tech hiring. According to the data, this has plunged 27% in IT overall, 32% in IT quality assurance and 23% in IT product management. Software engineering jobs (coding), which were long considered recession-proof, are down 26%. That kind of cutback in coders would have been unthinkable during the post-Covid boom era when programmers were treated like rare minerals – so scarce that they needed to be preserved at all costs, regardless of how the economy was faring.
Source: LinkedIn data
As a PhD researcher, I was constantly reminded that correlation does not equal causation, but I have a horrible feeling these early declines in white-collar hiring are a nasty harbinger of worse to come. If I think of my own use of AI, it has already replaced my front-end developer for certain projects. A few of my fellow entrepreneurs keep pestering me to recommend an AI marketing assistant so that they can reduce costs, and a straw poll among 10 or so of my freelance creative friends reveals that it’s really tough out there. From translation to TV production, companies (mine included), are already looking at how to leverage AI to do more, or at least the same, with less staff.
For many industries in the automation firing line, the driver for this cost-cutting is survival rather than just increasing revenue. Take the example of Intuit, the giant US software company that sells the world’s most popular accounting software, QuickBooks. On Tuesday, its stock price dropped after The Washington Post reported that Donald Trump’s Department of Government Efficiency is discussing creating a mobile app for Americans to file their taxes with the Internal Revenue Service for free. Throw in the impending release of AI agents (mini versions of specialist AIs trained to do a particular task, combined with infinite knowledge of context) and it’s clear that many white-collar jobs will be automated away.
The IMF has been reporting on this for a few years now and, after modelling various automation scenarios, is now increasing its calls for a government response to what it terms the “AI job shock”. However, they propose the same old tired solutions: job support, reskilling, etc. in a somewhat disingenuously titled paper: “How Fiscal Policy Can Help Broaden the Gains of AI to Humanity”.
Even if the risk of automation is overplayed, the wider problem is wage deflation. Though many white-collar jobs will likely be retained due to political and social pressure.
Last week, I caught myself doing something that made me cringe. There I was, at 11 p.m., crafting what I hoped would be a ‘thoughtful’ comment on a connection’s post about AI. Why? Because I hadn’t engaged on LinkedIn for a few days, and the nagging voice in my head reminded me that ‘visibility is credibility’. It’s a bit like Facebook in the 90s when if you didn’t have a profile, you had serious credibility issues.
That’s when it hit me: LinkedIn has become the corporate equivalent of homework – something we do, not because we want to, but because we feel we must. And I’m not alone in this compulsion to engage in digital performance art. A recent study by Originality.ai, an AI content detector (no conflict of interest there!), confirms what many of us suspected: over half of LinkedIn’s long-form posts are now AI-generated, with a staggering 189% increase in AI content following the launch of ChatGPT.
It’s a massive pain to constantly have to ‘feed the content beast’ while juggling the competing demands of your job, family life and the expectation to maintain an ‘always-on’ digital presence. So, it’s not surprising that many are deploying AI tools to cheat their way to an active LinkedIn profile. Need to come across as an industry expert? There’s an AI for that. Want to comment on the latest tech trends? AI can whip up a response in seconds. A whole industry has sprung up to write LinkedIn spam (like this?!) so you don’t have to.
Come to think of it, using AI also seems to be the entirely appropriate response to the cynical monetisation of our attention and contacts by Microsoft (the owners of LinkedIn). At any one time, my home page runs the gamut from ‘inspiring’ stories about how people have overcome serious illness to get back to work, to people bigging up their new roles and companies.
But what can replace it given that, despite its flaws, recruiters still scroll through profiles, business deals still (apparently) spark from LinkedIn messages, and industry insights (even AI-generated ones) still shape professional conversations? I’m not entirely sure, as any replacement will suffer from a lack of numbers to drive the network effect, and most people don’t want to have to use multiple apps to do what they can currently with a single one. However, if all the content on LinkedIn is AI crap and recruitment is done via algorithms, people will abandon the platform and seek smaller, more authentic trust-based networks that meet a specific need.
For now, we probably all have little choice but to remain on LinkedIn, but we should rage against the algorithms by focusing on quality over quantity, genuine engagement over performance metrics, and posting useful insight over corporate spam.
In the UK it’s the time of year when schools get to the end of their curriculums and set 16 year-olds on the final leg of study before taking their GCSEs in the early summer. In the Yahaya household, I’m struggling to get my son to buckle down early enough to avoid the last-minute stress of cramming all he’s learned in two years into a few desperate weeks of late-nighters.
So I’m trialling the below process, using a patchwork of AI tools, to bolster his exam prep:
Collect study materials and create a NotebookLM
Tell the AI to structure and prioritise content that most often appears on past papers
Generate practice questions to benchmark where your child is
Create a study plan based on the outcome of the test
Use AI for tutoring
Reinforce learning
Closer to the time, simulate final exams
Quick revision
Working linearly through a traditional textbook is fundamentally dull so I recommend trying the above to mix it up and use AI at its best.
Just when you thought you could at least get a job at your local factory, BMW drops a video of its progress on using humanoid robots at its Spartanburg facility. What sets this video apart from the scary but ultimately pointless ones of acrobatic robots is that these robots are actually useful. They can autonomously lift, transport and place battery components onto car chassis with precision. Figure, the company that produces these robots, claims that, since their introduction in August 2024, its humanoid robots have achieved a 400% increase in the speed with which electric car batteries can be assembled and a sevenfold improvement in accuracy.
This marks (yet another) pivotal shift in the use of automation in manufacturing, with profound implications for the future of physical work in any industry that involves picking stuff up, assembling it or manufacturing it. It’s worth reading the blurb on Figure’s website to see where this is all (quickly) going.
As automation continues to integrate with human life at scale, we can predict that the labor-based economy as we know it will transform. Robots that can think, learn, reason, and interact with their environments will eventually be capable of performing tasks better than humans. Today, manual labor compensation is the primary driver of goods and services prices, accounting for ~50% of global GDP (~$42 trillion/yr), but as these robots “join the workforce,” everywhere from factories to farmland, the cost of labor will decrease until it becomes equivalent to the price of renting a robot, facilitating a long-term, holistic reduction in costs. Over time, humans could leave the loop altogether as robots become capable of building other robots — driving prices down even more. This will change our productivity in exciting ways. Manual labor could become optional and higher production could bring an abundance of affordable goods and services, creating the potential for more wealth for everyone. Figure CEO Brett Adcock
Wow, so eventually it will be cheaper to rent a robot to replace the global labour force that currently ‘costs’ 50% of global GDP. Just as with digital service delivery and AI, even if most factories will continue to be manned with humans, there will be a constantly ruthless background calculus of whether a machine can do the job better, cheaper and faster.
That’s all for this week. Subscribe for the latest innovations and developments with AI.
So you don’t miss a thing, follow our TikTok, Spotify, YouTube, Instagram and X pages for more creative content and insights into our work and what we do.
Network Hub, 300 Kensal Road, London, W10 5BE, UK
We’ll send you quarterly updates so you can see what we’re working on, what the future holds and how we’re shaping it.
