A $48 Million Paint Job: What the Reflecting Pool Teaches Us About Leadership

Stop Painting the Pool: What I told the 70th Commonwealth Leaders Forum last month, and what the data told me first.

Walking down London's Pall Mall to the vaunted address of Marlborough House last month, I wondered whether to temper myself. This was, after all, the 70th anniversary gathering of the Duke of Edinburgh's Commonwealth Study Conferences: current and future leaders from business, politics and civil society across the 56 Commonwealth countries. Or I could be blunt about the future I believe we are facing. The sun was blazing and a heatwave was about to set fire to London anyway. I decided a little more heat could only get people fired up.

Twenty-five years of watching leaders across the world has taught me one thing. In times of crisis, the leaders who help us survive are the ones who ask the questions no one else will, listen bravely, and lead with no regard for personal popularity. And be in no doubt, the world is currently in a cascade of crises.

Are we developing the wrong kind of leaders? That was the blunt question posed to me and my fellow panellists.

My answer? Just look at the Lincoln Memorial Reflecting Pool in Washington DC.

The Reflecting Pool has now been renovated twice: at a cost of $34 million in 2013, and $14 million more in 2026. This time the money went on painting it "American Flag Blue" for the 250th anniversary of the United States. Within days of both renovations, the green algae the bureaucracy was trying to remove was back.

The reason is simple. The conditions that produce the algae were never fixed: the shallow, almost still water, DC's humidity and heat, the pool's insufficient filtration system. Washington's bureaucrats kept fixing the aesthetics without changing the system itself.

This is the model for how power responds to structural problems it does not want to solve, or does not know how to.

The morning of the panel, the UK announced a ban on social media for under-16s, following Australia and France, and under pressure from 116,000 public consultation responses. The government called the move "world-leading". And let's be fair, banning access is not nothing. But these are the same platforms proven to build the radicalisation pipeline. Their engagement algorithms, the business model that monetises the attention of children: none of it is being dismantled or penalised. Australia is already proving there are ways around its ban. Which makes the ban simply,  a blue coat of paint. The architecture of harm underneath remains entirely intact.

We never fix structural problems because we are always chasing the next exciting story. The next technology, the next investment thesis, the next headline that promises value and wealth, without looking at the deeper consequences already forming underneath. That means the people doing the actual structural work are operating without the infrastructure they need, because the infrastructure budget is going on paint.

The answer is not that we are developing the wrong leaders. It is that we are refusing to rebuild the systems that would let the right ones function.

Leadership in the age of visibility

The question is one every leader across the globe is grappling with. How do you lead in an age of visibility? You can convolute it, root your answer in fear, or you can do it bravely. You make it make sense.

In business, the leaders who have earned trust did it by standing for something at cost, not by managing the story. Think of Rose Marcario, the former CEO of Patagonia, or Whitney Wolfe Herd at Bumble.

In geopolitics, the most consequential leaders operate under impossible scrutiny and ignore it entirely, because the stakes are too high for performance. New Zealand's former Prime Minister Jacinda Ardern and Finland's former PM Sanna Marin are strong examples of this approach.

But for the majority in politics and in business, values and actions are increasingly shaped by the algorithm. Be it the assumption of what people want, or the fear of what the mob will punish.

The leaders who will last are the ones who understand that scrutiny is not a threat to manage. It is the only reliable signal that the work is real.

The trust deficit

Trust between people is broken because we have always cherry-picked who gets heard. That is not a new phenomenon, but technology has exacerbated it. We have masked who is collating the data. We have put its computation in a black box. And we have ignored the lens of those bearing the actual cost of our regional decisions. The confrontational stance of West versus East, North versus South was not created by social media. Social media has simply laid it bare for all to see.

The evidence shows just how bare. The Western male's grievances have spread globally via the manosphere to geographies where women have not made even half the gains it purports to fight against. Meanwhile in the G7, the Reykjavik Index for Leadership shows declining trust in women leaders among young people, the first reversal in a generation.

The most capable AI models being sold across the globe are trained on the distorted record of mostly Western data, already riddled with bias, before we even get to the lack of representation. These models are now handing their blind spots back as objective output, with no author and no analytics to argue with.

This is no longer a communications problem to be solved with good storytelling, though that would help a little, if rooted in facts, figures and honesty. Leaders today face a legitimacy reckoning. And the institutions that retain trust will be the ones that did the work: calculating the ripple effects of their decisions and internalising the voices of those who understand not just the crux of the issue, but have working knowledge of solutions pertinent to their region. Especially from the geographies that drive the world's agenda but have never set it.

 The next generation of influence

Influence has been weaponised. The same infrastructure that was supposed to surface new voices and democratise access to power has been turned into a delivery mechanism for suppression and manufactured consensus.

I keep coming back to the manosphere economy and its adjacent tradwife movement as an example. It now runs across India, Nigeria, South Korea, Japan and the West, built on a business model that monetises purposelessness in young men and the frustrations of young women. The platform algorithm suppresses expertise that does not fit the existing power structure. And we know this is deliberate, because fairness is a product the industry already sells to enterprise clients.

So how does the next generation use this moment to create real impact for good? In this editor's humble opinion, three things.

• First, see it clearly and break out of siloed thinking. Refuse the inherited frame, connect the dots others treat as separate, and update your picture faster than the world is changing.
• Second, build your solution differently. Do not just occupy the existing structures. Replace the infrastructure, without replicating what you tore down.
• Third, go where the energy is. Pay attention to the geographies being ignored, underfunded and misread. Those are the future, and where the real leverage sits. The next person who changes how power works is already operating somewhere the powerful have not looked yet.

 A view to 2035

First, look back at the chain reaction in Western economies over the past 15 years. Austerity hollowed out public services. Tax policies let technology companies walk away with the revenue that should have funded them. Growth slowed, inflation rose, education budgets were slashed, jobs disappeared. Young men with no economic future and no framework for understanding why were handed an answer by a manosphere economy that seemed to have been waiting for exactly this moment. Right-wing nationalism filled the political vacuum. And nobody connected the dots in time.

Now look at what we are building next. AI data centres can consume as much power at peak demand as the entire American state of New York. Stanford's AI Index estimates that training a single frontier model produced 72,816 tonnes of CO₂.

For what? Productivity, they say. But who will purchase the result of that productivity, if no one has the funds to? And if young people can’t enter the jobs market, how will any sector have senior staff in 10-15 years?

Stanford's payroll research shows employment for software developers aged 22 to 25 has fallen nearly 20 per cent from its late-2022 peak, with young workers across the most AI-exposed occupations down 16 per cent. A story that went viral in China is that of a Chinese actor fully booked in 2025 , who had to return to his village and sell vegetables in the market. Why? AI is now used to create the dramas he used to act in.

The compounding consequences are forming again, in real time, and once again nobody is joining the dots. A tool that could be trained to help humanity is now in danger of being stuck in a doom loop of simplistic debate: is it good, is it bad.

The Commonwealth Secretariat calls AI a democratising technology. A chance to level the global playing field. And the data backs the instinct, because emerging economies are already adopting AI faster and more confidently than the West. Oxford Insights identifies the 12-country Commonwealth AI Consortium as one of the only multilateral structures capable of building AI governance outside the US-China duopoly.

The question is no longer whether Commonwealth nations will use AI. They already are. The questions we should be asking are these. How will it empower, not disenfranchise, the workers it purports to support? And whose values, whose data and whose definition of benefit gets encoded into the tools they use? That answer is currently being decided in two countries, by a handful of powerful people. Not the 56 Commonwealth countries, nor the 195 officially recognised countries on this blue planet.

The takeaway

The next generation of leaders has a task no generation before them has faced: to decide whether AI becomes the technology that finally distributes power, or the most efficient tool ever built for concentrating it.

That decision cannot be made by the people building the models. It must be made by the people governing their use.

And the Commonwealth, home to 2.7 billion people, with more geographies, more languages and more lived experience of what happens when technology is built for someone else, has more power than any other body on earth to rewrite the brief. If its leaders choose.