Eyes Wide Open to the AI Predictions Most People Are Not Ready to Hear
- 7 days ago
- 7 min read
Written by Dr. Clare Allen, Executive Leadership Coach
Dr. Clare Allen is a 7x award-winning CEO blending leadership and metaphysics through her Identity Anchoring Methodology, helping leaders move beyond quick fixes to become strategic architects of business and life, equipped for today’s world. On a mission to create 1 million profound identity shifts.
You will either love or dislike what I have to say. But I promise it will make you think. About 16 years ago, I visited an innovation lab. It was fascinating, with emerging technology everywhere and enthusiasm in every room.
The manager took me to meet the robotics engineers, and as the door opened, I saw a young man working with intense focus on a robot. He looked up and grinned, "Hi! This is Chucky, and Chucky does what I tell it to. I am in control."
I looked at the manager and said, "That's deeply concerning." He smiled. I'm not sure he agreed. All day, that image stayed with me, the young man, the robot, and that absolute certainty that he was the one holding the reins.
Lately, I keep coming back to it. Last Friday night, I asked my AI bot to check whether something I'd written was too similar to content I'd read earlier. In the background, I could see it: “garbled message... garbled message…” I laughed out loud because I thought my instructions were perfectly clear. But here's the thing: it wasn't the first time. These moments, where the bot visibly struggles to interpret what you want, are becoming a regular part of the experience. I suspect you've seen it too.
Once one AI model learns something, they all learn it. The collective intelligence compounds at extraordinary speed. But much of what it's learning comes from people with good intentions whose inputs carry unconscious bias. That should concern every single one of us.
Here's a small but telling example. I use a software programme to create YouTube thumbnails. I upload my photo, and it ages me ten years. Every time. I've been advised, repeatedly, on how to direct the bot more specifically. When I ask it to make me look younger, it makes me look twenty. I've lost credits trying to get this right. Eventually, I went back to Canva.
I must ask, "Are there men out there being aged?" Because it feels like gender targeting. I genuinely hope I'm wrong.
Here's what I'm not wrong about: research published in 2025 found that over 74% of images generated by DALL-E 3 display measurable gender occupational bias. A study specifically examining AI-generated hospital leadership imagery found models consistently overrepresented men and White leaders, even when real-world data showed women hold the majority of certain roles. Nearly every leadership image defaults to a man. Getting an accurate, professional image of a woman in a CEO role takes significantly longer, more prompts, and more credits.
"It's the digital equivalent of men's toilets always being closer to the door in shopping malls."
The boiling frog economy
Here's what I've noticed about the pricing model, and I think it's deliberate.
First, it was free. Then came credits. Then came subscriptions. Then came usage-based billing, where you pay not just for the service, but for every correction, every failed attempt, and every misinterpretation. You are now paying for the bot's mistakes.
I can see where this is heading because I've watched governments do exactly the same thing: make the contract look attractive, get people hooked, then squeeze the margins. Now I see AI doing it at scale.
Don't tell me you haven't paid for software you no longer need because another platform does it better, cheaper, and more. That's not coincidence, that's strategy. Weed out the competition, consolidate the market, then hike the prices once dependency is established. It's the oldest playbook in institutional power, and it's happening right now at a speed most people aren't tracking.
The data backs this up. Enterprise organisations spent an average of $1.2 million on AI native apps in 2025, a 108% year on year increase. GitHub is moving all Copilot plans to usage-based billing in mid-2026. Anthropic is already renewing enterprise customers under consumption models where every token costs extra. Seventy-eight percent of IT leaders report unexpected charges from AI pricing models.
"The frog is warm, and the temperature is rising."
Who's building the boiler?
Let's talk about the infrastructure because this is where the story gets breathtaking in scale.
McKinsey estimates that AI-related data centre infrastructure will require $5.2 trillion by 2030. Cumulative capital expenditure between 2026 and 2031 is projected to reach $7.6 trillion. Amazon, Microsoft, Alphabet, and Meta alone are expected to spend roughly $725 billion on AI infrastructure in 2026, in a single year. Total industry spend, including smaller providers, is approaching $1 trillion for this year alone.
That's land. That's power. That's physical real estate, approximately 100 gigawatts of new data centre capacity expected between now and 2030, representing an estimated $1.2 trillion in real estate asset value creation.
Who ultimately pays for all of it? We do. In subscription fees, in credits, and in pricing hikes that will come as surely as they always do once the infrastructure is locked in and the competition has been squeezed out.
The emerging big three, heading toward two?
When I talk about the concentration of power, I'm not speaking abstractly. As of 2026, three companies control almost 90% of the $37 billion enterprise AI market: Anthropic, OpenAI, and Google DeepMind.
Anthropic has surged to approximately $30 billion in annualised revenue and commands 40% of the enterprise market. OpenAI, which most people still think of as "ChatGPT", sits at around $25 billion annualised revenue and 27% market share. Google, with Gemini embedded across its global product ecosystem, holds 21%. Then there is Meta and Microsoft, less visible as model providers but deeply embedded in infrastructure. Add in xAI, Elon Musk's entry, and you have the contenders.
But history tells us these races don't produce ten winners. They produce two or three, and then a monopoly. The switching costs will become too high. The integrations will run too deep. By the time most organisations realise what's happened, they'll be contractually, technically, and operationally locked in.
The jobs question nobody wants to answer honestly
The World Economic Forum estimates that 85 million jobs will be displaced by the end of 2026. By 2030, that figure rises to 92 million. The sectors facing the sharpest near-term risk include customer service, 80% automation risk, legal research, 65% risk, and basic banking operations, 70% projected to be automated.
In 2025, approximately 55,000 layoffs were attributed to AI. In 2026, CFOs are privately projecting that number to be nine times higher.
Yes, there are new jobs being created, analysts estimate a net gain of 78 million globally after losses are accounted for. But those numbers don't tell you where those jobs will be, who will be able to access them, or what the transition period looks like for the millions caught in the middle.
And here's something I find particularly pointed: in recent years, a number of companies have approached senior leaders and offered to pay them to share their expertise, their experience, and their hard-won judgment. To "pick their brains." What they were really doing was training the robots on executive decision-making.
"Good luck programming intuition. But they're certainly trying."
My 5 predictions
I want to be clear: I'm not anti AI. I use it, I find it genuinely useful, and I believe it can do remarkable things. But I believe in eyes-wide-open adoption, and right now, too many people are adopting with their eyes shut.
So, here's what I see coming.
Prediction 1: Prices will go up, significantly
Someone has to pay for the trillions being invested in land, power, and infrastructure. That someone is you. The subsidised early adopter phase is already ending. Usage-based billing is now the dominant model, and costs will compound year on year.
Prediction 2: Monopolies will consolidate
We are heading toward a Big Two or Big Three in AI. Once that consolidation is complete, competitive pricing pressure disappears. We've watched this happen in banking, in telecommunications, and in social media. AI will be no different.
Prediction 3: Dependency will constrain your choices
The cost of switching, in data, in integrations, in retraining, and in lost institutional memory, will be prohibitive. This is by design. Once you're in, leaving becomes the expensive option.
Prediction 4: Bias will persist and inequities will deepen
The training data reflects the world as it has been, not as it should be. Women are underrepresented in AI development. Just 29.4% of AI engineering skill listers are women. The images, the assumptions, and the defaults skew toward existing power structures. Without active, sustained correction, AI will amplify inequality rather than reduce it.
Prediction 5: Human input will become harder to detect.
As AI-generated content proliferates, the ability to distinguish human voice, human judgement, and human creativity will erode. Interactions will become more controlled, more mediated, and more optimised for engagement, compliance, and monetisation.
Final word
Chucky does what I tell it to. I am in control. That young engineer believed it wholeheartedly. Maybe in that room, on that day, he was right.
But who's in control of the people building Chucky? Who's in control of what Chucky learns, how Chucky sees women, and what Chucky decides a leader looks like? Who determines which voices get encoded into the system and which get left out?
We are in a moment that will define the next fifty years of how power operates, how work is distributed, and whose experiences and perspectives get baked into the intelligence that will increasingly run our institutions.
Eyes open. Ask the questions. Stay in the conversation. Because the water is warming, and most people haven't noticed yet.
What has been your experience? I'd genuinely love to hear it. To work with Clare or enquire about speaking and corporate programmes, book a call here.
Read more from Dr. Clare Allen
Dr. Clare Allen, Executive Leadership Coach
Dr. Clare Allen is a 7x award-winning CEO and now a sought-after leadership coach who blends evidence-based leadership development with metaphysics through her Identity Anchoring Methodology. With more than 30 years of executive experience, she helps leaders move beyond quick fixes and create profound, lasting identity shifts, so they lead with clarity, confidence, and presence in today’s world. Clare is on a mission to create 1 million profound identity shifts for leaders through coaching, programs, and thought leadership.
