Paris Diary: As AI Energy Demands Test Cities, Dublin Offers Key Lessons

22 02 2025 | 13:21 Special to The Energy Mix

Are you an Artificial Intelligence optimist or pessimist?

That seemed to be a popular question at the AI and Democracy panel during the AI Action Summit in Paris. It was amusing to hear a panel of Stanford professors declaring themselves. When I headed out to Paris, I was torn. Now that I’m returning home, I feel less torn, perhaps more optimistic and highly skeptical.

AI is definitely here and it will become deeply entrenched in all aspects of a city’s functioning. But it brings with it many serious issues that need to be carefully considered, planned for, and monitored.

Which brings me to a focus on AI and the impact it will have on energy. Cities spend thousands each year on energy procurement. Some cities operate their own municipal utilities, and consequently have some control over energy production, distribution, and costing. In British Columbia, most cities rely on our two major utilities – BC Hydro and Fortis.

A city with a population of one million will generally spend 12% of its budget on energy to power buildings, water and waste facilities, fleets, public amenities like swimming pools, and other services. It is difficult to estimate how much this would increase as AI started to displace operations and services, but national reports provide some insight.

According to Berkeley Labs, data centres in the United States consumed 70 billion kilowatt hours (kWh) of electricity in 2014, representing 1.8% of total U.S. consumption. By 2023, it jumped to 4.4%, and with significant advances in AI, it could rise as high as 12% by 2028.

There is no doubt that the demand for electricity in our cities is growing. As zero-emissions mandates drive the adoption of electric vehicles, as extreme weather creates greater need for heat pumps and air conditioners, as homeowners, businesses, and industrial processes rely on AI features to operate and serve their clients, and as cryptocurrency that relies on data mining proliferates, the competition for stable and affordable electricity will increase.

This competition for electricity has driven some jurisdictions to restrict the construction of data centres. In B.C., there is a temporary moratorium on allowing BC Hydro to service new connection requests for cryptocurrency mining projects due to their 24-hour, high-energy operations and the strain they impose on the electricity grid. In Dublin, Google’s plan to build a data centre was denied and the city imposed a moratorium until 2028 on all new data centres for similar reasons—officials were concerned that the facilities would draw too much power away from the electricity grid.

It’s ironic that Dublin should place a moratorium on new data centres because at the AI Action Summit, that city was recognized as a global leader in AI. According to the Dublin Economic Monitor, “Dublin is poised to seize the mantle of Europe’s AI capital. Industry projections paint an ambitious picture: by 2030, AI could contribute a staggering €15 billion to Dublin’s economy, representing 12% of the city’s projected GDP.”

Dublin’s approach demonstrates both the optimistic and pessimistic sides of AI. It is charting a future vision for the city that supports AI innovation in human resources, financial monitoring, smart energy, urban transportation, and medical technology, while placing guardrails to protect its energy, people, and environment.

As a member of the EU, Dublin is governed by the EU AI Act, which advances regulations to support the sustainable and ethical use of AI, a measure the current U.S. administration labels protectionist. The city council is also trying to support companies that develop sustainable products to reduce greenhouse gas emissions and support work force development. But I wonder how council deliberations will change as AI becomes a more prominent feature of the economic landscape and the pressure for access to electricity increases.

It’s early days, but I think there is a great deal Canadian cities can learn from Dublin’s experience. It’s exciting to see the new innovations that have emerged. It’s also instructive to see the thoughtful approach the city has taken to the ethics, sustainability, and governance of AI.

If cities can deploy AI to better and more affordably meet the needs of residents, if they can ensure it is smart, sustainable, and inclusive and delivered in an ethical, safe, and socially just manner, then I’m an optimist and I can embrace its adoption.

But, that requires mayors, councils, and city staff having their eyes wide open to the risks and trade-offs that AI presents: risks related to governance, redirecting vital energy resources away from public needs, exposing city data to potential bad actors, and creating vulnerabilities in the operation and delivery of city services. But with careful planning, partnerships across public, private, and non-profit organizations, and clear social and environmental objectives, cities can mitigate their exposure and gain the benefits of AI.

Shauna Sylvester is the Founder and Lead Convenor of Urban Climate Leadership, a project of MakeWay that works with cities to support their efforts to create healthy, safe, and resilient communities. Shauna will be sharing her learnings live from Paris, especially as they relate to AI, cities, and energy.

AI Governance, Democracy Have Big Implications for Cities

Two nights ago, Canada’s Ambassador to France, Stéphane Dion, hosted a gathering of Canadians who are attending the AI Action Summit in Paris. It was a full room of mostly private sector companies with some prominent Canadian AI academics, civil servants, and NGOs working on AI governance and democracy.

Prime Minister Justin Trudeau held bilateral meetings in an adjoining room for several hours while the rest of us enjoyed canapes and wine provided by the Alliance Français des Industries du Numérique.

As I walked back to the Métro station along cobblestone streets, I was struck by a couple of impressions from the event.

• The first was seeing Team Canada’s efforts to diversify our Canadian economy in full swing. Clearly, our government officials were working their European networks to support investment in Canadian AI companies and profile them to new markets.

• The second was how out of time it all felt. In the United States, it seems that a coup d’étât is nearly complete, with unelected computer engineers now replacing most of the major government databases and treasury and service delivery systems with new, proprietary artificial intelligence systems. But that theme wasn’t on the agenda.

As I returned home and compared notes with my colleague Mairin Loewen, I was somewhat comforted to learn that the AI conversations taking place in Paris are broad ranging. Participating Canadians are organizing and speaking on a range of issues—from AI and sustainability, to AI and participatory democracy, to AI and Indigenous sovereignty. Representatives from at least five Canadian universities (McGill, Simon Fraser University. Toronto Metropolitan, University of Alberta, and University of Toronto) are helping to raise awareness about the scope and adoption of AI in Canada and shape a broader international conversation on AI governance.

Prime Minister Trudeau’s public statements also reflect this broader conversation. In his remarks Monday, he noted that AI must be in service to all—not just oligarchs—and it must drive economic growth and equal benefits, including enhanced access for the middle class and for emerging economies globally.

The Emergence of AI for Cities

Where do cities fit into this discussion?

It’s clear in our research that AI is an emerging issue for cities. Some cities have adopted AI in their service delivery. But it is early days for most municipal governments that are just coming to grips with what AI is, and how it can help them meet the needs of their local communities. Few municipalities have developed AI plans, and fewer still have considered the governance of AI or its impact on their current and future energy needs.

So what are the most relevant AI applications for cities?

In cities globally, AI has been deployed successfully in smart energy management. AI can optimize energy distribution through smart grids, support energy efficiency, forecast energy demand through predictive analytics, and integrate renewables like solar and wind—enabling energy to be produced where it is consumed.

AI has also been incorporated into traffic management systems to reduce congestion and improve commute times. It’s driving autonomous cars and powering smart transit systems. AI can enable multimodal, integrated transportation in cities and improve energy efficiency, decrease greenhouse gas emissions, and support greater access and capacity for people and goods movement.

I recently had the opportunity to serve on the Technical Expert Advisory Committee for the Earthshot Prize, where I reviewed dozens of nominations from cities around the world. A number of them profiled the use of AI to monitor air and water quality through surveillance cameras and worn sensors to detect and respond to pollution problems in real time. Communities were also using AI to monitor waste, improve recycling rates, identify and monitor the health of urban forests, and locate thermal leakage in buildings.

Several sessions at the AI Action Summit are probing uses of AI to better analyze data for public safety and emergency responses, to increase administrative efficiencies and speed up service delivery, and to proactively address climate change.

It’s incredible when you think of all the ways that AI can be used to improve operations, future forecasting, and service delivery within our cities. It’s even more incredible when you think of how fast the field is evolving. It really is moving at breakneck speed, which brings its own challenges.

How Vulnerable Might Our Cities Be?

In one session I attended, I learned that the frameworks for governing AI aren’t keeping pace with the evolution of the technologies. Most of the major companies like Google, Meta, Amazon, OpenAI, and Microsoft are committing to safety frameworks to ensure that their AI technologies do not increase the risk of biological, chemical, or nuclear warfare. But developing even these frameworks, which should be a basic minimum, is a huge undertaking.

And if big companies are finding it difficult to assess, monitor, and mediate the risks that their AI models present, imagine how vulnerable local governments might be if they implement AI systems without thinking through their use and governance.

In another session on building trust in the age of AI, I compiled a list of questions that cities might want to adopt as they begin to design their AI plans. These questions are about assessing a city’s readiness for AI.

• How do we ensure that the information cities collect isn’t compromised when new AI systems are implemented? This question was prompted by news reports this week of all the information that was lost when the U.S. government deleted websites containing important scientific, medical, and cultural information. It is also relates to existing privacy regulations and guidance for data/information integrity and archiving—so that we have the means to store important information and ensure it isn’t tampered with by bad actors.

• How do we educate citizens and build AI literacy? Cities can begin by targeting AI literacy efforts with civic employees (city staff, police forces, school boards, parks, or transit employees) so that they better understand the AI applications in their own work. Strong AI literacy programs can position residents and employees to identify risks and raise alarm bells when AI isn’t acting as intended in its application.

• How can cities create collaborative partnerships to support literacy, policy, and technology development? Cities are one part of an ecosystem that is involved in the design, development, and deployment of AI. Developing strong partnerships across private, academic, legal, non-profit, and other orders of government, early in the adoption of an AI plan, will bring much-needed intelligence, accountability, and support to cities.

• How can cities use their procurement power, incentives, and capacities to support good AI actors? Cities have power through their purchasing decisions, who and what they incentivize, and how they advocate. Cities can use that power to advance AI technologies and companies that are ethical leaders and block “bad actors” that don’t adhere to international protocols and standards.

While these questions don’t specifically consider AI and energy, they provide the baseline for thinking about AI governance.

In my next note from Paris, I’ll probe the energy issue more directly and look at the capacities and energy costs of AI.

Learning How AI Could Shape Cities’ Energy Needs

It’s 4 AM on Sunday morning. I’m on the 6th floor of a hotel in the heart of the Latin Quarter in Paris, managing my jet lag by contemplating the role of Artificial Intelligence (AI) for cities. Around me are the sounds of sirens and partiers whose shouts and hollers sometimes shift to a rising chorus of a song I don’t recognize.

This week, Paris is hosting the AI Action Summit, a gathering of world leaders focused on the responsible development and deployment of artificial intelligence. There are hundreds of side meetings and I’m here with my Saskatchewan-based colleague Mairin Loewen to learn more about AI and what it could mean for cities. We are especially interested in how AI could impact the energy needs of cities.

Part of my learning is coming to grips with AI technologies which are evolving at breakneck speed. I’m married to a computer engineer so I know a little. I use ChatGPT regularly and I’ve appreciated DeepSeek’s capacity to help me make sense of research questions.

But admittedly, when I read that Elon Musk’s coders are implementing new AI systems on U.S. Treasury computers to find efficiencies, I get extremely nervous. Suddenly 2001: A Space Odyssey’s HAL9000 comes to mind. “Open the pod bay doors, HAL.” “I’m sorry, Dave, I’m afraid I can’t do that!”

So, it is with some trepidation that I enter this world. I’m both excited about the awesome power AI can offer cities if directed wisely, and scared about the potential destruction it can cause when unchecked.

My first step in this discovery is to better understand the recent developments in AI. It’s an expanding and complex field, so admittedly I probed DeepSeek and Chat GPT to better understand its evolution. Here are the five most important recent developments these AI platforms highlighted:

1. Tools like OpenAI’s ChatGPT, the recently-released Chinese platform DeepSeek, and DALL·E have revolutionized how we interact with AI. ChatGPT and DeepSeek generate human-like text for tasks like writing and coding, while DALL·E creates images from text descriptions. These platforms are used in education, customer service, and creative industries.

2. DeepMind’s AlphaFold is a breakthrough platform for AI in health care. It predicts protein structures which can accelerate medical research and drug discoveries. Similarly, AI-powered diagnostic tools, like those from PathAI, are improving disease detection and treatment accuracy.

3. Microsoft’s Planetary Computer uses AI to analyze environmental data, helping scientists track deforestation, predict weather patterns, optimize renewable energy systems, and address climate change.

4. MidJourney and Stable Diffusion enable users to generate artwork from text prompts, and in music, OpenAI’s Jukebox creates original compositions, opening new avenues for creative expression for artists and creators.

5. While not a technology, initiatives like the EU AI Act and Google’s AI Principles are shaping how AI is developed and used. These frameworks address the ethics and application of AI and aim to align it with values of transparency and accessibility.

Which brings me back to Paris.

As much as AI might scare us and threaten our livelihoods, it is here and permeating every aspect of our lives. This is why the French government has taken a leadership role in gathering leaders from across sectors to consider AI’s impact, ethics, and governance. It is also why I find myself in a hotel in the Latin Quarter reflecting on the transformative power of AI while contemplating serious questions that cities need to consider as they adopt, implement, and govern it.

Cover photo: Photo downloaded from www.wimklerkx.nl under Creative Commons License Copyright