AI is a leadership challenge as well as a technological one

I’ve been metaphorically staring at a blank page for nearly a year, wondering how I can even start writing about my thoughts and observations on AI. The Board, the organisation, our members and the wider sector are looking to us to articulate our stance.

As someone who feels the burden of responsibility that comes with a leadership position like mine, I’ve been overwhelmed by the complexity of the subject and the lack of expertise I bring to the debate.

Something has changed recently. Not a sudden realisation of what the answer is, or a sudden confidence that I have something to say. The public launch of our strategy has given us a stake in the ground: clarity about what we stand for and our purpose to improve the conditions for Civil Society to thrive.

There was also Kate Lee’s instruction: “Nick, you need to write a blog on AI”.

So, like I’ve suddenly been pushed on stage, blinking into the lights, a thought hit me: I’m not alone. How many other leaders, trustees and CEOs are being looked to for answers without having the expertise or breadth of knowledge to make sense of all this?

But that’s leadership, right?

Leadership is about recognising a need and creating the conditions to address that need responsibly, whilst navigating uncertainty. There’s huge pressure to “roll out AI”, “respond to the dangers of AI”, “keep humans in the loop” or “write an AI policy”.

If we’re not running at 1000 miles per hour, “optimising”, “experimenting” and intervening in existential debates, it can feel like we’re failing. There’s a fear of being left behind and a fear of implementing.

It reminds me of those old Batman episodes where there seemed to be quicksand everywhere (I grew up thinking there were quicksand dangers behind most bushes). The instinct is to struggle to get out rather than calm down and make sense of the situation.

This isn’t a failure of leadership. It’s increasingly what leadership feels like.

So how do we make sense of it?

AI is already being used here, sometimes officially and sometimes not.

Colleagues are using it in various ways, often with real productivity benefits.

We’ve also encountered the risks of moving too quickly. Teams have explored unofficial tools without fully understanding the processes, controls and data standards required. Some solutions bought outside IT approval proved unreliable and difficult to support.

Our clearest lesson is that production AI needs much stronger governance than experimental AI. Documentation, testing, transparency and human oversight all matter.

That means questions bigger than the technology: How do we govern AI responsibly and build capability and trust? I believe there are fundamental human qualities we must protect; critical thinking, judgement, empathy, conflict resolution and belonging. And how do we avoid leaving people behind?

AI is becoming part of our societal fabric. I’ve toyed with various analogies for how this might play out, but the one that works best for me is electricity.

I think of AI as the power of capability rather than the capability itself. It helps me stay oriented on the essential considerations, including how we make sure we don’t electrocute ourselves!

Electricity started as a competitive advantage, became essential infrastructure, transformed almost everything, created dependency, required regulation and eventually became largely invisible.

Until it fails.

AI looks similar. Restrictions around access to frontier AI models and geopolitical tensions reinforce that AI is becoming part of national infrastructure, not simply a software product. That raises questions about resilience, sovereignty and access that charity leaders need to consider.

The point isn’t to provoke fear, but to encourage strategic thinking.

Economically, AI capability is currently subsidised. If the true costs were passed on to consumers, I guarantee we’d see fewer AI-generated images of Shrimp Jesus or Will Smith eating spaghetti.

But would we be able to give up the articulate board reports and on-tap productivity gains quite so easily?

As AI becomes embedded in organisational processes, what happens if the economics suddenly change? Will it get more expensive?

There’s also cognitive dependency. I’ve noticed a real difference between ideas I’ve developed myself, including those 4am brainwaves I mull over and elaborate on, versus concepts I’ve developed with AI.

There’s growing research about what AI dependency might mean for critical thinking, judgement and reasoning. As leaders, we need to think deliberately about protecting the human qualities we’ll depend upon in the future.

The greatest risk isn’t that AI starts thinking for us: it’s that we’ll stop thinking for ourselves.

This is where I would advocate for a positive outlook.

In the roundtables and discussions I’ve been part of, there is a danger of getting pulled into a vicious circle of doom and gloom when we have an opportunity to create a virtuous and positive response.

AI is here to stay. So, what would a future look like where we harness its benefits whilst minimising the harms? It could mean:

• every organisation has access to capabilities such as legal guidance, fundraising support, research, translation, communications and operational advice, allowing greater focus on mission and delivery.
• human relationships and qualities such as empathy, trust, care, community, conflict resolution, belonging and judgement are protected and valued.
• communities are better able to organise locally, with civil society enabling rather than delivering solutions.
• knowledge, research and evidence are more accessible, enabling communities to influence the decisions that affect them.
• nobody is left behind.

This gives us something to organise around.

What capabilities does the sector need to make this possible? What do we need in terms of leadership, people, governance, data, organisations, partnerships and learning? We can use these questions to think about maturity, risk, ownership and the collective action needed to bring a vision like this to life.

This is where NCVO can help.

We won’t be AI experts on everything. But we can help civil society build capability together: convening difficult conversations, turning experimentation and lived experience into practical learning, connecting charities with technology experts, government and academia, surfacing unintended consequences and translating technical developments into leadership and governance questions.

We can champion approaches that enable every organisation, regardless of size, to participate.

But we can’t do this alone.

We need leadership, honesty and all of you to bring your insights and experience to bear to create the conditions in which the right answers can emerge so we can steward ourselves toward a thriving and sustainable future.

Find out more about NCVO’s priorities for the future in our new strategy, including how we’re approaching AI and supporting civil society to navigate change.