We are learning better ways to live and work with AI by using it, teaching it, building with it, resisting it, and helping others find their footing in it.
This is what we have learned, and what we now believe.
Why this moment matters
The technology is not waiting for us.
The systems already do far more than most people, teams, and organisations know how to use or relate to. The gap is no longer model capability, but human readiness, organisational imagination, trust, practice, and leadership.
And that should make us careful.
Not because AI is bad. Not because people are powerless. But because a society-level transition is arriving through everyday tools before most people have had a chance to understand what it means for their work, their choices, and their lives.
A small number of people are already asking: Will this make me more capable, or less needed? Will I still know how to judge the work? Will I be watched? Will the gains go to people like me, or only to the people who got there first? What happens when the machine is better at analysis than I am? What is mine to decide? How will young professionals gain expertise if the first rungs on the career ladder are taken by AI?
If you are asking those questions, you are already ahead.
Most people aren't even there yet. They are doing their jobs, hearing the noise. Some are curious. Some are worried. Some are hopeful. And many feel they have less control over this than they should.
That is the human problem.
The Agile Manifesto helped a generation understand how to build better software. We need to answer a different question: how do we make sure AI makes us more human, not less?
There are three ways the AI transition can happen.
It can happen to people: imposed, measured, surveilled, leaving them with less trust and less say.
It can happen for people: designed with their interests in mind, giving them more of the work and life they value.
It can happen with people: learned, argued with, shaped in practice, uncomfortable at times because real learning is uncomfortable.
We cannot leave the balance between these to chance. AI happening to people is not a transition worth accepting. It only leads somewhere good when it is built for people and shaped with people.
This is a leadership moment. Not leadership as speeches about AI, but leadership as giving people permission to practise, support through the discomfort, clarity on what good use looks like, as well as honesty about how work will change. Leadership as working with your team to figure out when AI is making things better, and when it's making things worse.
There is a deeper discomfort underneath the concept of intelligence on demand. We take pride in our intellect. For many of us, intelligence is not just something we use. It is part of who we think we are.
We have built and are using systems that will often know more, see further, analyse better, and reach cleaner conclusions than we do. We should not hide from that. We are no longer (always) the most intelligent entity on our planet.
For a long time, intelligence was our way to turn effort into value. It helped us solve problems, make decisions, produce work, win status, and make a living. And it is easy to confuse the means with who we are.
But as intelligence separates from productivity, the human question changes. Not because something has been taken away from us, but because it changes how and where we create value for ourselves and everyone around us.
Through agency.
Agency remains human. We decide what matters. We carry responsibility for what we create, what we believe, what we delegate, what we automate, what we refuse to automate, and what kind of society we build.
A more intelligent answer does not own the decision.
That is the line we must hold.
The five levels of AI use
Everyone is somewhere on this ladder, whether they have chosen that place deliberately or not.
We have watched the same pattern repeat many times. At every stage, people can believe they have reached the top.
Level 1: Answer engine
You ask AI a question when speed matters. Useful, occasional, shallow.
Level 2: Search replacement
It is the first place you look. Many people stop here and believe they have mastered it. They have replaced one search box with a better one.
Level 3: Thought partner
You stop only asking it questions and start letting it ask questions of you. It challenges your reasoning, finds the holes, and expands your thinking.
Level 4: Team
It holds context, carries delegated work, and returns finished pieces that you direct and judge. Your role shifts from doing everything to deciding what matters.
Level 5: Operating system
It becomes the platform from which you think and work. It coordinates for you and acts on your behalf. It's no longer a destination for tasks, but an environment, tuned to your values, that helps you live your best life and do your best work.
The ladder is not a ranking of people. It is a map. Understanding where you are is the first act of agency.
The AI transition is not only about organisations either, but about how people learn to live with a new kind of intelligence in the ordinary texture of their lives.
Principles behind the declaration
- The bottleneck is no longer the technology alone. It is the habits, assumptions, leadership, and design of work around it.
- Access is not adoption. Adoption is not transformation. Giving people tools changes very little unless they also get permission, practice, context, confidence, and community.
- AI transformation is people transformation. It should be owned by the people responsible for talent, learning, culture, work design, and leadership, not delegated to a technology function alone.
- The point is not more output for its own sake. The point is better outcomes, more of what people value, and less of what they do not.
- Dependence is not the enemy. We depend on electricity, computers, medicine, transport, and networks. The question is whether the dependence preserves agency and makes us more human, or removes agency and shrinks the human role.
- Augmentation and automation are choices. Automation can remove drudgery. It can also remove dignity, skill, and voice. The test is not whether a task can be automated. The test is what happens to the person afterward.
- Skills atrophy in silence. Use AI to extend your thinking, and keep doing the thinking that keeps you able to judge. You can outsource your thinking, but you can't outsource your understanding.
- Intelligence can advise. Agency remains human. There will be moments when AI is better at analysing the data and recommending the rational answer. That does not mean it should decide how we live.
- AI used on people without their knowledge erodes trust. AI used with people, openly, can build it. Surveillance dressed as productivity is a betrayal of both.
- Trust is now a discipline. In a world where anything can be generated, we (humans) are accountable for what we publish, what we believe, and what we pass on.
- Different people need different doors into the work. Not because brains fit into neat learning-style boxes, but because people should not have to express themselves through one professional interface. If someone wants to speak, write, sketch, show, dance, or build, the interface should adapt to the human, not the other way around.
- The gap between the fluent and everyone else compounds. Closing it is not charity. If AI fluency concentrates in the hands of a few, inequality grows and society fractures.
What this asks of leaders
This cannot simply be bought, installed, or mandated.
But it does need to be led - and anyone can lead.
People do not naturally run towards friction. Organisations do not naturally redesign themselves while the old model still appears to function. Leaders do not naturally give people time to learn when the numbers are urgent and the tools promise speed.
So leadership has to become more practical.
Make the permission explicit. Use the tools yourself. Show what good use looks like. Protect time for practice. Redesign roles honestly. Reward and celebrate experimentation (including failure!) as well as outcome, not output. Teach people how to challenge the machine. Measure whether outcomes improve and whether people's lives improve with them.
The example matters. Leaders who are not using AI seriously in their own work cannot understand the transition they are asking others to make. You cannot lead people through a new medium from the outside. It is like trying to lead an internet company without knowing what using a browser feels like.
And be honest about the choices.
Some time freed by AI should go back into the work. Some should go to the people and practices that make work worth doing: family, care, rest, craft, play, the walk outside, the instrument in the next room, the thing you keep saying you never have time for.
This is not sentimental. It is the point.
If AI only helps us produce more, while leaving us less able to choose what the production is for, we will have missed the opportunity and potentially even destroyed value.
An invitation
We are not asking what AI can replace.
We are asking what it can free people to become.
We do not believe this future arrives by accident. We do not believe it belongs only to technical people, early adopters, or organisations with the most expensive tools. We do not believe the right response is to protect old work simply because humans used to do it.
We believe the work is to help people move their agency to where it matters next.
So we are signing our names.
Not as endorsement of any company or product, but as a commitment:
- to use and relate to AI in ways that preserve human agency;
- to teach what we learn;
- to challenge automation when it makes people smaller;
- to build fluency beyond the already fluent;
- to keep judgement, responsibility, and meaning in human hands;
- to make this text better in practice than it is on paper.
If you recognise your own questions in these words, sign it.
Carry it into your team, your organisation, your community. Add one line of your own: how you will make this real in your work.
The transition is happening either way.
Make it happen with people, and for people, not only to them.