Artificial Intelligence and Circular Economy: opportunities and risks

By Dr. Isabela da Cruz Bonatto

That day has come: let the one who has never used artificial intelligence cast the first stone. It’s no longer something out of a futuristic movie or a sci-fi episode. In fact, episodes of Black Mirror have been unfolding before our very eyes—in real time. I’d even say we’re going through an identity crisis: humans are discovering they’re not as irreplaceable as they once thought, and that they can be easily substituted by combinations faster than neurons.

I’ll confess: the other day I used some AI platforms to conduct a benchmarking study on electronic waste management (ah, the irony). In just one hour I had a solid base of global information—something that would’ve taken me at least a week to build before. Interconnected, accessible studies and databases, capable of helping us find better, faster and smarter solutions to complex environmental problems—there’s no denying that.

But that very same power can be used for both good and bad. Google itself recently disclosed the environmental impacts behind seemingly harmless searches: every time you ask Gemini a question, it consumes the same amount of energy as watching nine seconds of TV. A single query emits 0.03 grams of carbon dioxide and uses about five drops of water. That might sound small, but on a global scale… the effect is massive. So much so that Google’s emissions have increased 51% since 2019 because of AI. And it’s not just Google. The International Energy Agency estimates that electricity demand from data centers is set to more than double by 2030, reaching 945 terawatt-hours—more than Japan’s entire annual consumption. On the other hand, here’s some good news: an August report showed that systems are becoming more efficient—the energy needed to run a medium-sized text prompt is now 33 times lower than it was 12 months ago. So yes, demand is rising, but there are also clear advances in energy and material efficiency (or so we hope).

The relationship between AI, the environment and sustainability is no longer in question; the real question now is: how does it actually translate into progress for the Circular Economy?

A recent study published in ScienceDirect shows that AI is becoming a driving force for circular business models. Capabilities like predictive analytics, real-time monitoring and intelligent automation help redesign production chains to regenerate, reuse and repurpose—almost as if the algorithm were a circular architect. But there are risks: without solid circularity indicators, the promise can turn into a mirage. We need clear metrics to monitor the life cycle of products and materials and ensure that AI is truly closing loops—not just optimizing the linear model. In real life, that means having the right indicators on usage, returns, reuse, waste attention and product life cycle—and trusting that the algorithms are providing the right diagnosis. Not everything is technological sunshine and roses.

Another interesting angle comes from a study by the Ellen MacArthur Foundation with support from McKinsey: they show that AI can accelerate circularity in three main areas—design, new business models, and infrastructure optimization. Translated into everyday life: AI could help create packaging that disassembles itself at the end of its life cycle, support leasing systems that extend product lifespan, and even improve reverse logistics to recover and recycle everything we consume. The gains are tangible: up to US $127 billion per year in food and US $90 billion per year in electronics by 2030. We’re talking about real money being saved and recycled—in a system that learns and adapts. In other words, digitalized circularity is also about competitiveness and profitability—which makes it even more irresistible in a capitalist world.

And let’s bring in Harvard Business Review to back this up: according to Shirley Lu and George Serafeim, the world remains stuck in a linear take-make-waste cycle, even though circularity promises trillions in value. But it hits barriers like the low value of used products, high separation costs and lack of traceability. The way out? Accelerate with AI in three very practical fronts: extend product lifespans, use less raw material, and increase the share of recycled content. AI can help maintain high useful life with updates (as with iPhones) or through product-as-a-service strategies, where the company remains the owner and the consumer just “rents”—effectively extending the real usage cycle. That turns into revenue, builds loyalty, increases the value of used products and pushes a more circular and profitable economy—as long as the technology doesn’t become just another expensive luxury.

This is where we need to connect the dots.

Circular Economy teaches us to rethink material and energy flows, pursue efficiency, eliminate waste and regenerate systems. But when it comes to AI, we face a paradox: it can accelerate solutions and opportunities for circularity (like mapping flows, forecasting recycling chains, optimizing reverse logistics, identifying waste hotspots or even speeding up research into new materials), but it can also amplify environmental impacts if not used consciously. Some of the key risks include AI’s environmental footprint (with its growing energy and water consumption in data centers), e-waste (the race for chips, servers and supermachines also generates mountains of electronic waste and pressures critical-mineral mining), and digital inequality (developing countries may become dependent on expensive technologies without fair access to their benefits).

The big challenge is balance. We need AI in service of circularity—not the other way around. How can we ensure that Artificial Intelligence, instead of worsening the environmental crisis, becomes an effective part of the solution? We must keep our critical thinking sharp. We can’t just be carried away by the tech hype. It’s time to choose: do we want AI that deepens inequality and environmental pressure, or AI that boosts the transition to a circular economy?

I try to stay optimistic. I believe processes will become increasingly efficient, with lower energy consumption and better resource use. What today seems like a dilemma—more AI meaning more energy demand—could reach equilibrium in the future, as long as the same creativity used to write algorithms is applied to reducing impacts and regenerating systems. We can use AI as a strategic ally of circularity—with sharp eyes and solid criteria: demanding efficiency, traceability and transparent metrics.

True intelligence is not measured only in lines of code or processing speed. In the environmental arena, only circularity will ensure that this intelligence is real—and not merely artificial. In the end, the real challenge won’t just be about creating and monitoring artificial intelligence… but about building circular intelligence.

*Dr. Isabela da Cruz Bonatto is an ambassador for the Circular Movement, holding a Ph.D. and a Master’s degree in Environmental Engineering from the Federal University of Santa Catarina, with an MBA in Environmental Management. She works as a socio-environmental consultant, focusing on solid waste management, promoting the Circular Economy, and corporate sustainability. Residing in Kenya since 2021, she is a board member of the Together for Better Foundation and works directly with NGOs to combat period poverty and develop sustainable waste management solutions.

*This text was automatically translated with the help of artificial intelligence and reviewed. Still, there may be slight differences compared to the original version in Portuguese.