Hi First name / friend—
Yesterday was Day One of the AIFOD Summit at the United Nations in Geneva, and I’m still processing everything that happened.
Even getting into the building is an experience. We had to register with the United Nations about four weeks in advance and be approved before attending. Yesterday morning we showed our ID, went through security and received our digital badges for the week.
Then we walked into the General Assembly Hall.
I’m not going to pretend I wasn’t a little awestruck. It’s that room you’ve seen on television countless times, except suddenly I was sitting in it as part of a conversation about the future of artificial intelligence and international AI policy.
That was a pretty surreal moment.
One comment from the opening sessions immediately stuck with me:
A seat at the table isn’t enough. You need a voice in the room.
That pretty much summed up why I came here.
I wasn’t invited to Geneva just to sit quietly and say I had attended something at the UN. I came because small businesses need to be represented in these conversations, and because I’ve been talking about responsible and ethical AI for a long time.
My goal for the entire week was simple. If I could contribute one idea that might actually make a difference, I’d consider the trip worthwhile.
Apparently, I didn’t need the whole week.
Can small businesses really compete in AI?
One of the biggest discussions yesterday was whether artificial intelligence is becoming concentrated in the hands of a relatively small number of companies and countries.
The answer in the room was pretty clear.
Yes.
That creates some uncomfortable questions around AI sovereignty.
Who owns the infrastructure? Who controls access to the models? What happens when businesses, communities or even countries become dependent on technology controlled somewhere else?
This isn’t only a developing-country problem.
Canada needs to think about this too.
Every country should.
Because if we don’t build this properly now, we risk becoming dependent on infrastructure and decisions we don’t control.
One of my favourite comments from yesterday was essentially this:
If we don’t build properly, we’ll end up renting our future.
That one stayed with me.
We also talked about whether small companies can realistically participate when AI requires so much money, computing infrastructure, data and talent.
The biggest disadvantage for small businesses is pretty obvious: cost.
Large organizations have deeper pockets.
But small businesses have something they often don’t.
Agility.
If I want to change direction in my business, I don’t need a board meeting, three committees, six approvals and twenty people climbing up an organizational ladder before somebody finally says yes.
Small companies can test something, realize it isn’t working, change direction and try again.
That ability to pivot quickly is going to matter tremendously as AI continues changing.
Then we started talking about the environmental cost of AI
This is where the conversation really grabbed me.
I already came into Geneva believing that education needs to be part of responsible AI adoption.
I believe that even more now.
Most people use AI without seeing what is happening behind the screen.
You type something into ChatGPT or another AI system, get an answer back almost instantly, and it feels incredibly simple.
But there is infrastructure behind that.
- There are data centres.
- There is electricity.
- There is cooling.
- There are environmental costs.
There are also questions around what information we’re feeding these systems, where that information goes and who ultimately controls it.
Right now, AI feels a little like giving candy to a kid and sending them loose in a toy store.
We’ve basically handed the public an incredibly powerful technology and said, “Have fun.”
Meanwhile, most people haven’t been taught much about the consequences of using it.
We’ve spent decades teaching people things like turning lights off, unplugging appliances and conserving electricity.
Maybe AI literacy needs to include the same kind of thinking.
Responsible AI shouldn’t only be about what information we put into the system.
It should include how we use the technology itself.
Should the biggest AI users pay more?
Another part of the environmental discussion involved the amount of energy being consumed by large AI and data-centre operations.
The idea of charging very large users more came up.
And I agree with it.
Traditionally, the more electricity or resources a large company purchases, the better pricing they can often negotiate.
But if an organization is placing an extraordinary strain on infrastructure and creating a disproportionate environmental impact, maybe that model needs to change.
Why should the biggest users receive the biggest discounts?
And more importantly, why should residential users and small commercial businesses eventually absorb part of the infrastructure cost needed to support massive AI operations?
If you use dramatically more, perhaps you should pay dramatically more.
That seems pretty reasonable to me.
Then somebody said something that changed how I was thinking about the problem
One of the panelists, from Romania, talked about democratizing infrastructure and potentially using smaller, more localized data centres instead of continually concentrating everything into enormous facilities.
And that immediately triggered another thought for me.
My husband mines cryptocurrency from computers running in our basement.
We paid for the equipment and electricity, the computers contributed computing power to a larger network, and he received tokens in return.
It was essentially a distributed computing model with an economic incentive attached to it.
Then my brain went somewhere else.
Torrents.
With distributed file-sharing, information doesn't necessarily sit in one giant central location. Different pieces can exist across multiple computers and locations while still functioning together as part of the same network.
So I asked the panel a question.
Could we apply some of that same thinking to AI infrastructure?
Instead of assuming the future has to be dominated by enormous centralized data centres, could some AI computing or storage be distributed across smaller locations?
Could those locations use solar or other renewable energy?
- Could smaller communities participate?
- Could a remote region, for example, have one solar-powered facility serving several surrounding communities?
Maybe the facility becomes more than a data centre.
Maybe it becomes a place where people access AI, education, information and technology.
Now, I’m a bookkeeper and business consultant.
I’m certainly not walking into the United Nations claiming that I’ve suddenly solved global data-centre architecture.
So I put the idea to the experts in the room and asked whether it was actually plausible.
The moderator started writing.
At the end of the session, three ideas were selected to move forward into the drafting discussion.
Mine was one of them.
Later that evening, during the networking reception, another attendee came over and told me he had been impressed by the suggestion.
That was probably the moment yesterday when everything really hit me.
I came to Geneva hoping I might contribute something useful.
On Day One, I did.
Small doesn’t mean powerless
If there’s one thing I took away from yesterday, it’s this:
Being small doesn’t automatically put us at a disadvantage.
Small businesses don’t have the financial resources of multinational corporations.
Small countries don’t have the infrastructure of the world’s biggest economies.
But being small can also mean fewer layers, faster decisions, different perspectives and an ability to challenge assumptions that larger organizations sometimes accept as inevitable.
Maybe the future of AI doesn’t have to mean bigger and bigger systems controlled by fewer and fewer organizations.
Maybe some of the answers come from thinking smaller.
More distributed.
More local.
More intentional.
There are still two very full days of conversations ahead, and eventually the work here will contribute to the Geneva Compact being developed for the next stage of discussion in Nairobi in 2027.
I’m going to keep sharing what I’m seeing, what concerns me, what gives me hope and where I think small business needs to push harder.
Because we shouldn’t simply have a seat at the table.
We need to make sure our voice is actually heard.
Day 2 is now starting, keep an eye out for my Day 2 thoughts.
Tanya
