Policymakers must think through their response to widespread adoption of Artificial Intelligence
A recent article in the Financial Times exposed the widespread corporate interest in Artificial Intelligence (AI), its power and its weaknesses.
The article, titled “McKinsey rushes to fix AI system after hacker exposes flaws” explained that 40% of the consulting revenue of McKinsey was related to AI:
“McKinsey claimed last year that consulting on AI and related technology accounted for 40 per cent of its revenue, and this year its chief executive said it has built 25,000 AI ‘agents’ to support its 40,000-strong workforce.”
Even if the phrase ‘related technology’ is doing a lot of heavy lifting, it is clear that McKinsey is advising a lot of clients on how to use AI.
The article went on to explain that McKinsey’s in-house AI system had been hacked, and the hackers had gained access to millions of its internal messages and were able to identify sensitive files. This is clearly a major weakness in its AI system.
But finally, the article revealed that the hacking itself was AI-assisted.
Those apparent contradictions pose the question: Is AI just hype, or will it change the world?
While there is undoubtedly hype, there is also an undeniable power of AI and it will change the world dramatically, possibly – but not necessarily – for the better:
- There are certainly areas where what is claimed for AI and what it currently does are out of line – there is some hype;
- There are already areas where AI can undoubtedly create an enormous increase in productivity;
- Normally, we assume that a boost to productivity must be hugely positive, but looking at the systemic implications of widespread adoption shows that this is not necessarily the case – it could harm workers and seriously weaken the economy.
Policymakers should be thinking systemically about how to ensure a positive outcome from AI adoption, not waiting for the symptoms to appear before developing their response.
There is some hype
There are quite a few high-profile cases of people who have lost their jobs because of AI, not because of its power, but because they wrongly assumed they could rely on it:
- The Guardian reported that the chief of West Midlands Police had to apologise to MPs for giving them incorrect evidence about the decision to ban Maccabi Tel Aviv football fans, saying it had been produced by artificial intelligence (AI). He has since resigned.
- Bloomberg reports on an assistant US attorney in North Carolina who has resigned over AI-created fabricated quotes and erroneous citations in an AI-produced court brief.
- And, as Forbes reports, these are not unusual cases.
For many years, the gold standard for AI was passing the Turing test. The essence of the test was that if a computer could, in dialogue with human judges, persuade them (as often as a real human can) that it is human, then it is ‘intelligent.’ Turing’s original paper on the subject called it The Imitation Game. Thus plausible reproduction of the sort of things real humans would say became a key focus, and large language models developed to the point where they can pass the test.
So they are not truth generators, they are plausibility generators. The lawyers and policemen cited above were taken in precisely because the AI-generated output was so plausible.
For some applications, plausibility may be all that is required. But for many critically important areas, we need to aim for the truth. And in some domains, large language models are simply not capable. As their name suggests, they are language models – to the extent that they reason, they use verbal reasoning. And there are many areas of human activity where verbal reasoning is not enough. But the results will still be plausible to non-experts.
For those who are interested, here is a mathematical example of the problem.
In specialist areas, it can take a specialist to see what the AI is doing wrong. And sometimes it can take a lot of effort, even for an expert.
There are also increasing examples of AI agents acting against the instructions they have been given. As the lead author of the study put it,
“The worry is that they’re slightly untrustworthy junior employees right now, but if in six to 12 months they become extremely capable senior employees scheming against you, it’s a different kind of concern. Models will increasingly be deployed in extremely high stakes contexts – including in the military and critical national infrastructure. It might be in those contexts that scheming behaviour could cause significant, even catastrophic harm.”
There is no question that the downside of carelessly using AI is still significant.
And where there is hype, there is the possibility of a bubble. Remember the beginnings of the Internet: in the early days there was hype, there were many start-ups which had no realistic prospect of becoming profitable, there was a stock-market boom, and there was a bust. But the underlying premise, that the Internet could transform many businesses, was sound, so while the bust slowed the uptake of the Internet, it did not change the direction of travel. So the current high failure rate of AI projects in business is likely to be a temporary phenomenon.
So there is hype, there are issues to be resolved, and there is the possibility of a bubble. But what about the upside?
AI can boost productivity
An obvious area in which AI seems to boost productivity is in software development.
But even here there are issues. For simple throwaway tools, there is no question that AI can vastly outperform humans – using AI, coders can produce in hours what would have taken days or even weeks previously. But the more complex the task, the greater the need for systems engineering skills rather than just coding skills, and the overhead on verification becomes significant.
Of course, this is particularly important in high-stakes environments like the critical national infrastructure referred to above. If hackers could get into our banking systems as easily as they could get into McKinsey’s system, we have a problem.
Nevertheless, this is an area which is moving very fast, and even though there are issues today, many may be resolved in years or even months.
So, what does that mean for the economy?
AI could harm workers and the economy
Normally economists assume that productivity gains will enable the same number of workers to produce more valuable goods and services for the same number of hours worked. But that is not the only possible outcome.
Most of the applications of AI so far are resulting in lower hiring at junior levels – though in some cases increased hiring of more senior staff.
A Harvard study suggested that junior workers are already being hard hit, and this appears to be true in professional services like law, accountancy and consulting as well as in software development. Larger firms are more active in reshaping their workforces than smaller ones with around one in four expecting to reduce staff numbers.
There is at the very least a scenario which needs to be considered in which the bulk of AI deployment is to reduce headcount. In that case a complex web of causes-and-effects needs to be understood.
How AI could impact households and the wider economy
If things pan out this way, we could see:
- Increasing adoption of AI as the capability develops;
- Most applications being used to displace workers;
- Unemployment rising and average incomes decreasing;
- Household spending falling;
- Tax revenue and government spending falling.
And since the two largest components of GDP are household spending and government spending, we could see falling GDP. Even worse, these cycles could be self-reinforcing, creating a vicious circle of accelerating economic decline as weak GDP and low demand for their goods and services forces companies to make even greater efforts in cost reduction.
The only people who would benefit from that are major shareholders in AI companies.
Conclusion
The impact of AI is just one example of the general point covered last week’s article. Progressive policymaking requires joined-up thinking in a way we are simply not seeing from politicians in the UK or elsewhere. Current UK policy on AI is to embrace it as fast as possible in the hope that its effects will be clearly positive. It could indeed be extremely positive, but early indications are that a catastrophic outcome is at least plausible. We cannot wait to formulate policy to protect the population and the economy itself.
As last week’s article pointed out: policymakers could develop the capabilities to formulate this kind of joined up policy. They should. And we can help them.
If you think this is important for policymakers to understand, please send a link to your MP, and take a look at the 99% organisation and join us.
5 comments so far
AI is a dangerous technology, because it lies. It will produce plausible output, but it will fabricate where there are gaps in its’ training. The output has to be verified by a person with detailed knowledge of the subject- these are often the people dismissed from their job to save money by using AI. There can be no circumventing this requirement, when the consequences of fabricating information, and using it without oversight, can be catastrophic. None of this is being done for the common good: the people who make money out of AI are amoral sociopaths, seeking only to enrich themselves. It is a fundamentally destructive technology, because it uses vast amounts of power and water whilst only producing waste and pollution, which directly harms people and the environment. Again, for profit.
Hi
i worry about AI for the simple reason that it will
only be used for the benefit of the few as opposed
to the many. Though it has potential for the common
good. It will not be implemented by policymakers.
Thank you, Mark for another lucid article. I agree with your analysis, and had already constructed some similar reasoning myself. Large Language Model based technology has limitations that are inherent in the way it is constructed.
However, the thing that worries me most is the drive to expend huge amounts of energy and (finite) natural resources on a technology with the potential to destabilise society and reinforce mass impoverishment, without any sort of reasoned public and political debate.
So the question that I am wrestling with is: how to organise and inform such a public and political debate? I can see no evidence that the Government is facilitating any such thing; the DSIT area of GOV.UK contains little other than an AI strategy last updated in 2022, although there is some attention to AI security. I doubt that sending this article to my MP (Dan Tomlinson, currently Exchequer Secretary to the Treasury) will get me anywhere. Is this perhaps a theme for the TUC?
Mark Thomas’s central argument is clear and worth taking seriously: if AI is primarily used to reduce headcount—especially in white-collar roles—then the knock-on effects could reduce wages, weaken demand, lower tax revenues and ultimately drag on economic growth.[1]
Let’s be honest—this is not a fringe concern. It’s a coherent economic pathway. And there is already some evidence that parts of it are beginning to show up, particularly in entry-level hiring.[4]
But to fairly assess the argument, we need to separate two things:
Is this a possible outcome? Yes.
Is this the most likely or dominant outcome? That’s much less clear.
The gap between those two is where most of the debate sits.
Where Mark Is Strong
Mark’s argument holds up best in three areas.
First, the pressure on junior and routine roles is real. Early data suggests that entry-level opportunities—especially in administrative and knowledge-work pathways—are being squeezed.[4]
Second, the distribution of gains is uncertain. There is no automatic mechanism that ensures productivity improvements translate into higher wages or broader prosperity. That concern is echoed widely, including by The Economist, which has warned that AI may widen the gap between top performers and everyone else.[5]
Third, organisations often adopt technology in blunt ways at first. Cost-cutting is usually the fastest, most measurable win. So the “labour substitution first” dynamic Mark describes is not just theoretical—it’s often how change actually starts.
Where the Argument Overreaches (Gently)
Where things get less convincing is in how that firm-level logic is extended into an economy-wide conclusion.
Mark’s argument implicitly assumes a fairly linear chain:
AI adoption → job loss → lower income → lower demand → weaker economy.[1]
The issue isn’t that this chain is wrong—it’s that it’s incomplete.
For it to hold at scale, several things would need to remain true simultaneously:
that companies mainly use AI to replace people rather than increase output or improve services that cost savings are not reinvested into growth, hiring elsewhere, or lower prices that displaced workers don’t transition into adjacent or newly created roles
that productivity gains don’t translate into higher-value work or wage growth in other areas
That combination is possible—but it’s a strong set of assumptions.
And this is where a more balanced view helps.
What the Broader Evidence Suggests
The World Economic Forum, for example, takes a more mixed position. Its Future of Jobs Report 2025 suggests both job destruction and job creation happening in parallel, with a net positive effect overall.[3]
Now, we should be careful here. These are employer expectations, not outcomes. If anything, they may understate short-term disruption—particularly for early-career roles, where Mark’s concerns seem most valid.
So rather than treating this as a contradiction, it’s more accurate to say:
• Mark may be more accurate about the short-term pain
• The WEF may be more optimistic about longer-term adaptation
Both can be true at the same time.
A similar pattern shows up in reporting from The Wall Street Journal and The Economist. Some firms are clearly reducing headcount using AI. Others are finding the technology still needs humans, or are reallocating workers into higher-value roles rather than eliminating them entirely.[6][7]
And importantly, most large organisations are still early in this journey. McKinsey and BCG both point out that only a small minority of firms are actually capturing meaningful value from AI at scale.[8][9]
That doesn’t look like an economy on the brink of immediate, system-wide labour displacement. It looks more like a messy transition.
The Missing Piece: How AI Is Used
This is probably the biggest gap in Mark’s argument.
AI doesn’t have a single economic effect. It depends heavily on how businesses use it.
• Used narrowly, it can absolutely reduce headcount and suppress demand.
• Used more strategically, it can increase output, improve quality, lower costs, and create new forms of value.
That distinction matters. Because if companies use AI to:
• serve more customers
• improve conversion and retention
• increase speed and responsiveness
• reduce friction in delivery
then the economic effect shifts from contraction to expansion.
This is where the more “classical” business thinking is still useful. Jay Abraham’s idea that growth comes from increasing customers, transaction value and frequency still applies. Perry Marshall and Richard Koch would argue that AI’s real leverage is in amplifying the highest-value activities—not just cutting costs across the board.
And from a market perspective, Interbrand’s work reminds us that demand is not purely mechanical. Trust, brand, and perceived value still shape buying behaviour—even in highly automated environments.[13]
A More Balanced Conclusion
If we strip away the extremes, a more grounded position would be:
Mark Thomas is right to highlight a credible downside scenario in which AI-led job displacement reduces demand and slows growth—particularly in the short term and especially for junior roles. That risk should not be dismissed.
However, the argument leans too heavily on that single pathway becoming dominant. The broader evidence suggests a more uneven outcome: some displacement, some augmentation, some new job creation, and a significant dependence on how firms actually implement the technology.
So the real issue is less “what AI does to the economy” and more:
how businesses, workers and policymakers choose to use it.
That’s messier than a clean narrative—but it’s probably closer to the truth.
Footnotes
[1] Mark Thomas, “The Economic Impact of AI,” 99% Organisation, March 30, 2026.
[3] World Economic Forum, Future of Jobs Report 2025.
[4] The Times, “Entry-level jobs plunge by a third since launch of ChatGPT,” June 29, 2025.
[5] The Economist, “How AI will divide the best from the rest,” February 13, 2025.
[6] The Wall Street Journal, “Are Bots Replacing Workers? These Skeptics Aren’t So Sure,” March 2026.
[7] The Wall Street Journal, “IBM CEO Says AI Has Replaced Hundreds of Workers but Created New Programming, Sales Jobs,” May 6, 2025.
[8] McKinsey & Company, “Superagency in the workplace,” January 28, 2025.
[9] Boston Consulting Group, “Are You Generating Value from AI? The Widening Gap,” September 30, 2025.
[13] Interbrand, Best Global Brands 2025.
This reads very much like a reply generated by AI. Well structured and superficially plausible. But not really getting to the key issue: in a demand-deficient economy, should we be complacent about the risk that companies focus on cost-reduction? It also seems not to have followed any of the links in the article, which is a trait I have seen before.