Hi First name / friend—
If you missed my Day 1 recap from Geneva, you can catch up here first:
[Day 1 recap]. It’ll give you the background on where these conversations started and some of the thoughts I was already carrying into Day 2.
There was one moment during the opening remarks on Day 2 that really stayed with me.
We were sitting in a room at the United Nations in Geneva where Nelson Mandela and Mahatma Gandhi had spoken. That hit me harder than I expected it to.
These were people who stood in that space talking about humanity, independence, dignity and human rights. Decades later, we were sitting in the same room having many of the same conversations, just with a very different set of challenges in front of us.
How do we protect human dignity as technology changes the way we work? How do people stay relevant and continue to contribute? How do we make sure jobs evolve without stripping away purpose, judgment or value? How do we keep human responsibility in the decisions being made? And through all of it, how do we make sure we don’t lose the humanity in the name of progress?
It was incredibly humbling.
It also put the conversations of the day into perspective for me. AI may be new technology, but many of the questions we are asking are not new at all. Who has power? Who gets access? Who gets represented? Who gets left behind? And what responsibility do those with more power have to those with less?
That was probably the real theme of Day 2 for me.
The sessions covered everything from national sovereignty and negotiating with Big Tech to education, data ownership, infrastructure, AI agents, human accountability and collaboration between smaller nations. There was a lot to take in, and I’ll get into the more technical pieces in the briefing I’m putting together after the summit.
What I want to share here is what I personally walked away thinking about.
Sovereignty is a much bigger conversation than I realized
I came to Geneva already believing strongly in economic sovereignty. I don’t think any country should have to be completely beholden to another country for something that is becoming fundamental to its economy, education, government and future.
At the same time, I don’t think sovereignty means every country has to build everything itself. That’s not realistic, particularly for smaller or developing nations. One of the discussions today framed sovereignty more around control: who controls the data, who sets the rules, who makes the decisions and whether a country can change direction or providers without losing what it has built.
That last part stuck with me.
If you can’t leave, are you actually sovereign?
The easiest way I can explain what I mean is to think about renting an apartment. You rent the space. Fair enough. You know the building isn’t yours.
Then you move in your furniture. You buy a television. You build a bookshelf. You bring in your pictures, files and everything else you own. Maybe you even improve the space while you’re there. Then you decide to move and find out that everything you brought into the apartment has to stay behind.
Even the things you bought.
Would you consider that ownership?
That’s where my head went during some of the conversations about AI systems and AI agents.
We are putting huge amounts of information into these tools. We’re creating prompts, processes and workflows. We’re teaching them how our businesses operate. We’re adding our own context and knowledge to make them more useful.
And a lot of people are doing this using “free” AI.
Nothing in life is really free.
So what exactly are we giving in exchange?
One of the afternoon sessions went into the importance of portability. If a business or government decides to change AI providers, it shouldn’t just be able to download its basic data. The discussion included the accumulated workflows, configurations and operational knowledge that have been built inside the system over time.
That makes sense to me.
If I spend years teaching an AI system about my business and then have to start from scratch somewhere else because none of that learning can come with me, I wasn’t really building an asset I owned. I was improving somebody else’s rental property.
That applies to countries, but it applies just as much to small businesses.
Access and capability are not the same thing
The education sessions really got me thinking too.
One statement that stood out was that performance is not the same as understanding.
That is something I think we need to pay much more attention to. I can help someone produce a better paper, report, analysis or answer. That doesn’t automatically mean they understand it better. Can they explain why the answer is right? Can they tell when it’s wrong? Can they do the work themselves if they have to? Do they understand enough about the subject to challenge what the AI gives them?
Because if the answer to those questions becomes no, we have a problem.
One of the notes I wrote to myself was simply: remain independent thinkers while using AI.
That is probably going to become one of my biggest messages coming out of Geneva.
We need to teach people how to question AI, not just how to prompt it.
Verify the output. Ask why. Ask what might be missing. Ask whose perspective is represented. Ask what information the system was trained on. And sometimes, quite frankly, tell it that it’s wrong.
I’ve become more convinced that AI literacy is not about knowing which buttons to push.
It’s about knowing enough to challenge the machine.
The conversations about culture and language made this even more important. Much of today’s AI has been shaped by information and perspectives coming from dominant parts of the world. That doesn’t necessarily translate perfectly into another culture, language or community.
Knowledge does not automatically equal context.
We also have generations of human experience behind us. People have made mistakes, learned from them, improved things and passed that knowledge along. We need to be careful that, in our rush toward the newest answer, we don’t accidentally discard the experience that helped us get here.
AI can be an enabler or a divider
I wrote that sentence in my notebook today.
AI could be a divider, or it could be an enabler. I really think it depends on the choices we make now.
If only wealthy countries and large corporations can access the best infrastructure, the divide grows. If smaller nations become dependent on systems they cannot understand, challenge or leave, the divide grows. If people use AI without developing the critical-thinking skills to evaluate what it tells them, the divide grows there too.
But that isn’t the only possible outcome.
We heard a lot about smaller countries working together, sharing knowledge, collaborating on infrastructure and using their individual strengths collectively. They may not have the scale of the largest nations, but that doesn’t mean they have nothing to contribute. In some cases, being smaller can actually make them more agile.
I kept thinking about small business when I heard that.
Small businesses don’t have the resources of Microsoft, Google or Amazon. We’re never going to win by trying to become them.
We don’t need to.
We have different strengths. We can make decisions faster. We can experiment. We can collaborate. We can specialize. We can respond to our customers and communities without working through twelve departments first.
That agility has value.
Small businesses also understand implementation in a very practical way. It is one thing to talk about what technology could do. It is quite another to figure out how somebody with six employees, limited cash and a full workload is actually supposed to use it responsibly.
That voice needs to be part of the conversation too.
I originally came to Geneva because I wanted small business represented in these discussions. After Day 2, I think there is just as much for small businesses to contribute as there is for us to learn.
Every major sovereignty question being asked about countries can be brought down to a business level with a slightly different spin:
Who owns our information and what happens to it?
Can we move to another provider without losing everything we have built?
Are we developing human capability or creating dependency?
Who remains accountable when AI starts taking actions instead of simply giving answers?
Are we teaching our people enough to recognize when the AI is wrong?
Those are not future questions anymore.
What keeps coming back to me is humanity
For all the technology we discussed today, what I keep coming back to tonight is the room where we were having the conversation.
Mandela. Gandhi. Human rights. Independence. Dignity.
And now AI.
There were ministers, policymakers, educators, technology people, people from developing nations, representatives of smaller countries and people with very different life experiences sitting together trying to figure out what this next chapter should look like.
The developing world needs to be part of this.
Small nations need to be part of it.
Different languages and cultures need to be part of it.
Small businesses need to be part of it.
And the people who are going to live and work with the consequences of these decisions need to be part of it.
AI that excludes huge parts of humanity because they don’t have enough money, infrastructure or influence is not the AI future I want to see.
I believe AI should be for all.
That doesn’t mean everyone gets exactly the same technology or that every country has to build its own version of everything. It means we need to think intentionally about access, education, ownership, accountability and the ability for people and countries to maintain some control over their own future.
We need AI to enhance what humans can do without slowly taking away our ability to think for ourselves.
We need collaboration without dependence.
We need innovation without handing over every decision.
And we need to remember that the point of all this technology is supposed to be improving human lives.
Sitting in that room today, with all of its history, made that feel a lot less theoretical.
The technology may be artificial.
The consequences are very human.
Getting ready for the final day and reviewing the draft we are presenting for AI International Policy review.
Stay tuned…
Tanya
