Let's cut through the 'gówno prawda' (one of my favourties and thank you to my Polish ex) for a second. The conversation about AI and jobs has become a polarised mess of sales hype on one side and intellectual arrogance on the other. Both camps are wrong, and this stubborn refusal to see reality is costing businesses—especially in markets like the UK, where productivity growth has been all but stagnant for years.
Recent evidence shows just how gullible some business leaders have become. According to a survey by software platform Orgvue of more than 1,000 business leaders, more than half of those who laid off employees because they thought AI would replace their jobs are now regretting the decision. These companies fell for the marketing hype instead of understanding what AI can actually do today.
The disconnect between AI company promises and current capabilities is massive. But here's what the AI sceptics miss: this doesn't mean the technology is useless. It means we need to stop believing every marketing claim and start understanding what AI can actually do right now—which is still pretty impressive when used correctly.
On the flip side, we have the intellectual arrogance crowd who dismiss AI entirely because it's not perfect. Newsflash: humans aren't perfect either. We make mistakes constantly, and we do it at a much slower, more expensive rate than AI systems.
When I ask ChatGPT or Claude (my favourite) for help and it occasionally fabricates a quote, that's a limitation I need to work around. But when a human employee spends three hours researching something that takes AI three minutes to compile, even with verification time, that's a productivity problem that can't be ignored.
The key difference is cost and speed. Yes, AI makes mistakes. But humans make mistakes too, and they cost significantly more per hour to employ. This is where the "human in the loop" approach becomes essential—not just because AI is sometimes unreliable, but because combining AI speed with human judgement gives you the best of both worlds.
Let's look at real data instead of hypothetical scenarios. Amazon's fulfillment centers provide concrete evidence of what's happening today. Between 2022 and 2024, while package delivery volume jumped 20%, employment patterns revealed the truth about automation's current impact.
At delivery stations—where packages get loaded onto trucks—employment grew proportionally with volume. But at Amazon's most automated fulfillment centres, employment decreased by roughly 25% when accounting for productivity growth. That's not a future prediction; that's happening right now.
Here's what this means for your business: parts of jobs are already replaceable, and pretending otherwise is just willful ignorance. The question isn't whether AI will impact work—it's whether you'll adapt intelligently or get left behind by competitors who do.
Research from consultancy firm Gartner suggests that companies frequently cite "AI efficiencies" to justify financial decisions without demonstrating real productivity improvements. However, critics of AI adoption overlook a key finding from the same research: fewer than half of employees are actually using AI tools in their work, with only 8% leveraging these technologies to boost their productivity.
This massive underutilisation represents a competitive opportunity that most businesses are completely missing. While everyone argues about whether AI will replace jobs, smart companies are quietly using it to get work done faster and more efficiently.
Modern AI can crack historical challenges like the Enigma code "in short order," BUT it still has limitations. But those limitations don't negate its usefulness—they define how to use it effectively.
For the 33 million small business owners in the United States, the 5.45 million in the UK and countless more globally, the AI debate isn't academic—it's about survival in increasingly competitive markets. Three barriers currently prevent widespread AI adoption, but they're shrinking rapidly.
First, the technology reliability issue. Yes, current AI systems for business applications remain limited and require human oversight. But they're improving constantly, and the companies that start learning now will be ready when the technology matures.
Second, the data security concerns. Legitimate worries about sharing information with tech giants shouldn't paralyse decision-making. The solution is understanding which applications require sensitive data and which don't, not avoiding AI entirely.
Third, the cost barrier. While enterprise AI implementations cost millions, practical AI tools for productivity enhancement are increasingly accessible. The real cost isn't adoption—it's falling behind competitors who embrace these tools while you're still debating their perfection.
Here's the reality that AI sceptics refuse to acknowledge: in markets with limited growth—particularly in the UK where productivity has all but stagnated—AI represents one of the few available levers for competitive advantage. When your market isn't expanding, the only way to grow is to become more efficient than your competition.
This isn't about replacing humans wholesale; it's about productivity multiplication. When one person can accomplish what previously required three, that's not job destruction—it's competitive necessity. The businesses that figure this out first will dominate their markets.
Companies like Klarna are already demonstrating this advantage, replacing hundreds of customer service workers with AI systems while maintaining comparable service quality. Meta is using AI to augment programmers' capabilities. JP Morgan is deploying AI for tasks previously done by expensive MBA graduates. These aren't future possibilities—they're current realities.
The Silicon Valley vision extends far beyond today's limitations. Companies like Mechanize have secured major funding for "the full automation of the economy." Demis Hassabis from Google DeepMind believes artificial general intelligence could arrive within five to ten years.
Whether these timelines prove accurate doesn't matter for your immediate decisions. What matters is that serious money and talent are pushing toward comprehensive automation. Elon Musk predicts widespread job displacement. Bill Gates suggests humans won't be needed for "most things." These aren't fringe opinions—they're mainstream tech leader positions backed by massive investment.
The competitive implication is clear: while you're debating AI's limitations, your competitors are learning to use it effectively. By the time you decide the technology is "ready," they'll have years of experience optimising their processes.
Stop the intellectual onanism (see what I did there) and start learning what AI can actually do. The most successful approach treats AI as a powerful tool requiring human judgement rather than an autonomous replacement. This human-in-the-loop methodology acknowledges AI's current limitations while capturing its productivity benefits.
For AI believers like myself, we need to be honest about limitations while firmly advocating for adoption. Yes, AI makes mistakes. Yes, it requires verification. But it also processes information faster, works continuously, and costs less per hour than human alternatives.
For AI sceptics, your intellectual arrogance is becoming a business liability. Start experimenting with AI tools in low-risk applications. Learn what works and what doesn't. Develop the judgement to know when to trust AI output and when to verify. Your competitors are doing this while you're still debating whether AI is "really" intelligent.
The future belongs to businesses that combine AI capabilities with human judgement effectively. This isn't about choosing between humans and machines—it's about creating hybrid workflows that leverage both.
The companies succeeding with AI today aren't replacing workers wholesale; they're augmenting human capabilities while maintaining oversight for critical decisions. They're being realistic about AI's current limitations while preparing for future capabilities.
Most importantly, they're learning and adapting now rather than waiting for perfect technology. I'm crass so I'll steal from the cinematic masterpiece American Pie, replace sex with business in the following quote
- Victoria 'Vicky': It's got to be completely perfect. I want the right time, the right moment, the right place.
- Jessica: Vicky, it's not a space shuttle launch, it's sex.
In competitive markets with limited growth, that adaptation mindset isn't optional, it's called survival.
Stop believing the sales hype, but also stop dismissing the reality of what AI can already accomplish. The technology isn't perfect, but again neither are humans, and the combination of AI speed with human judgement creates competitive advantages that pure human efforts simply can't match. You need to speak with the excellent Salopian Robin Davis at Assuritivity for his Star Trek Borg analogy.
Your choice is simple: start learning how to use AI effectively, or watch your competitors gain advantages you'll struggle to overcome. The technology will continue improving whether you participate or not—but your market position depends entirely on how quickly you adapt, or die.