Faster, better… Greener – Is sustainable AI possible?
In the rush to adopt AI across the data and insights community, are we ignoring the urgent question of its significant environmental impact? From the energy demands of large language models to the infrastructure supporting them, the carbon footprint of AI is growing. The insights industry is not exempt from scrutiny.
The path to sustainable AI looks different across every market, shaped by local infrastructure, regulation, and economic priorities. WIN members from Ghana, Peru, the Netherlands, Sweden, Italy, Latvia, and Norway share where the conversation stands in their markets and what, if anything, is being done about it.
When the economy comes first
Across many markets, the environmental cost of AI is not yet the headline concern. The conversation is dominated by economics, and understandably so. Ato Micah (Managing Principal, Maverick Research, Ghana) reports that in Ghana and across West Africa, “most discussions have been around either job losses or the impact on productivity.” The environmental footprint of AI has not yet entered that frame.
In Peru, Urpi Torrado (CEO, Datum) observes the same pattern, explaining that this extends to how organisations make decisions. “In Latin America, organisations are still primarily focused on improving productivity and making AI more accessible, so speed and cost often take priority over environmental impact.”
Ethics first, environment next
Where regulation and public debate around AI have gained traction, the focus has landed on ethics, transparency, and data protection, not sustainability. For Karin Nelsson (CEO, Demoskop, Sweden), “the debate is still at an early stage and quite uneven. It’s picking up speed now, but it lags behind the conversations about AI ethics, copyright, and disinformation.”
Arnis Kaktins (Managing Director & CEO, SKDS, Latvia) describes the entire topic as “post-modern, a loss of stable meaning”. In a space where even the definition of ‘sustainable AI’ is contested, that observation deserves attention.
Most members expect sustainability to move up the agenda… eventually. Especially as energy costs may force the issue, as Rado Fonda (Head of Research, SWG S.p.A., Italy) notes: “It is possible that in the future the issue of AI’s electricity consumption will become central, because the cost of energy in Italy is already high and Italy depends heavily on imports.”
Where the conversation has already started
Sweden and the Netherlands stand out as markets where both regulation and practice are already moving. Karin Nelsson notes that Sweden’s national AI strategy, adopted in February 2026, explicitly states that AI’s environmental impact must be minimised, though she admits there are few Sweden-specific regulations beyond the EU AI Act adding obligations, requiring large providers to report on energy and resource consumption. Even so, she is candid about a persistent blind spot: “the Swedish debate tends to fall into an ‘AI will save the planet’ narrative, while the physical side, electricity, water, hardware, is often underestimated.”
In the Netherlands, Maaike Zijderveld (Principal Data Consultant, Motivaction) sees both sides of the same coin: AI is “becoming a more serious public issue because of constraints in electricity infrastructure and physical spaceAI can also help reduce other forms of environmental burden in research”. Her team in the Netherlands is already training researchers in responsible practices to avoid unnecessary AI processing, to improve both quality and efficiency. Though she admits they remain “cautiously optimistic that future models will become more efficient per task, making that balance more favourable over time”.
The Industry’s Role and Its Limits
If sustainable AI is to become a reality, who is responsible for getting there? For much of the network, the answer is clear: not individual agencies.
Ole Petter Nyhaug (Partner, Opinion AS, Norway) is direct: “This is a larger issue which should not be subject to each individual’s or company’s preferences or considerations. It needs to be addressed by industry organisations, international bodies, governments, and the global AI companies themselves.” Rado Fonda echoes this precisely: “Our role is to study the opinions, behaviours, and recommendations of the public and businesses, and the resulting actions must be taken by others, including institutions, governing bodies, regulators.”
Karin Nelsson names three further roles that insights professionals should undertake, being guardians of transparency, standard-setters, and,crucially, steerers rather than stoppers. “I share the position of KTH researcher Francesca Larosa, we need to steer rather than halt AI development. Our role isn’t to be AI-sceptics on principle,it’s to bring the nuanced evidence that makes wise choices possible.”
Urpi Torrado meanwhile brings the focus back to practice, “while sustainability is an important emerging consideration, our immediate responsibility is to help clients use AI thoughtfully, by validating outputs, protecting data quality, and avoiding overreliance on automated decisions.”
Across the network, the picture is consistent, sustainability is not yet the primary lens through which AI is being evaluated, but the conditions that will make it one, such as rising energy costs, tightening regulation, and growing public scrutiny, are already forming. The insights industry has a distinctive role to play in gathering the evidence that shapes the policies in the correct direction.