November 2025. I found myself once again in Lisbon, Portugal. The distinctive melancholy of the cobblestone streets mixed with the humid breeze from the Atlantic created a familiar atmosphere. Yet, around the MEO Arena, the air vibrated with a different energy: the anticipation of the future.
Web Summit 2025, one of the world’s largest tech conferences, had gathered startups, investors, and tech giants from across the globe.
At Web Summit 2024, Microsoft emphasized that advancing AI required massive infrastructure—specifically, data centers and a robust power supply. They spoke of the “body” and “blood” needed to sustain the colossal brain of AI. One year later, however, they posed a new, perhaps more profound question to us.
“We have the tool called AI. Now, how do we diffuse it into every corner of society?”
While Qualcomm’s vision of agent-centric edge AI was impressive (read my thoughts on that here), the session that struck me most this year was Microsoft’s keynote. It was a gritty yet essential narrative about the three elements required for AI adoption: Technology, Talent, and Trust.
The Fastest Tech Adoption in History
Microsoft Vice Chair and President Brad Smith took the main stage and began by presenting the “AI Diffusion Report.” Thanks to their ownership of ChatGPT, Microsoft possesses an incredibly high-resolution map of how AI is being used globally.
The screen displayed a graph comparing the speed of tech adoption throughout human history. Electricity, radio, telephone, internet, smartphones—compared to these game-changing innovations, the adoption curve for AI is spiking at an unprecedented angle.
“The fastest start to tech adoption in history.” There is no exaggeration in these words. We are witnessing a moment where the world is being rewritten at a speed never seen before.
Unexpected Leaders and the Widening Gap
So, who is riding this wave most skillfully? Many might guess the United States or China. However, Microsoft’s “AI Adoption Leaderboard” showed the United Arab Emirates (UAE) in first place, followed by Singapore. Small nations with strong, state-led digitalization strategies are outpacing the giants.
Conversely, when this data is mapped globally, a stark “red” reality emerges.
“An AI divide has emerged.”
Russia, large parts of Africa, South America, and Southeast Asia are lagging. Unfortunately, Japan is not entirely “green” either. A vast void is spreading, particularly across the Global South, where people are unable to reap the benefits of AI.
Infrastructure is Not Enough: Why AI Isn’t Used
Here, Microsoft offered a crucial insight. It is natural to assume that the lack of AI adoption stems from a lack of electricity or internet access. Based on their 2024 argument (see 2024 report), infrastructure deficits seemed to be the culprit.
However, the reality is more complex. Out of a global population of 8.1 billion, 7.4 billion have electricity, and 5.5 billion have internet access. The infrastructure is, to a large extent, already there.
The problem is skills. Only 4.2 billion people possess digital skills. Consequently, 3.9 billion people are left behind—not because they lack access to the grid, but because they lack the ability to use the technology.
In other words, having a power outlet and Wi-Fi means nothing if you don’t know the “manners” and “methods” to utilize AI. This isn’t just a developing world issue.
The Urban-Rural Divide
This disparity occurs even within a single country. Let’s look at the data from Portugal, the host nation. While AI usage is high in urban areas like Lisbon, it drops significantly in rural regions. There is a clear positive correlation between urbanization rates and AI user share.
This phenomenon is happening in Washington D.C., and likely between Tokyo and rural Japan as well. Information accumulates in cities, empowering urban dwellers. Meanwhile, rural areas are left behind, widening the gap. It is an ironic reality that as technology evolves, geographical advantages and disadvantages are accelerated.
Language and Cultural Barriers
As CityNomix, I also couldn’t overlook the issue of language. Much of the internet is in English, and AI training data reflects this.
English makes up 46.0% of the web content used to train Large Language Models (LLMs). In contrast, despite having more native speakers than English, Spanish represents only 4.6% of training data.
This means AI tends to “think” with the logic, culture, and values of the English-speaking world. For users of other languages, AI might remain a somewhat distant entity. This is another invisible wall preventing AI from truly becoming local.
The 3 Pillars of AI Adoption: Technology, Talent, Trust
National disparities, urban-rural divides, language barriers. To overcome these walls and truly implement AI in society, Brad Smith presented three essential elements:
Technology, Talent, and Trust.
The Technology is already here. Tools like ChatGPT, Gemini, and Copilot are evolving daily.
Next is Talent. We need to cultivate people with the skills to use AI in every country and region. This is urgent to close the “3.9 billion gap” mentioned earlier.
The Biggest Hurdle: Trust
Finally, the most difficult and fundamental element is Trust. Even with great technology and skilled talent, if people do not trust AI, adoption will stall.
“I see, Trust.”
Listening to the keynote, I nodded in agreement. This resonates strongly with my personal observations.
Back in Japan, many people who don’t use AI aren’t just confused by how it works; they feel a vague anxiety. “Will it take my job?” “Will it give me false information?” “Is my data safe?”
Even when I show them how I use Copilot to streamline my work, and they see it’s possible, a psychological barrier remains before they start using it themselves.
Some major companies even ban Generative AI entirely due to security concerns. While understandable, in a world evolving by the second, this maximizes the “risk of doing nothing.”
No trust leads to no use. No use leads to no understanding. No understanding prevents trust from building. Breaking this negative loop is perhaps the biggest challenge for AI adoption.
CityNomix’s Takeaway
Microsoft’s conclusion was clear: we must cultivate talent while simultaneously earning trust in AI. This is a homework assignment for all of us.
Trust in AI isn’t built solely on technical safety. It comes from accumulating small success stories where people realize, “AI is not an enemy, but a partner that extends human potential.”
We must share these small wins—and even our failures—to ground AI from “unknown magic” to a “useful tool.”
Amidst the excitement of Lisbon, I felt the difficulty of rooting this technology in society. Through Photomo’s theme of “Walk, Shoot, Write,” I intend to continue sincerely recording this evolving relationship between humans and technology.
