On an early October morning in 2025, two tourists took a walk to Sully Island, a small island a few hundred metres off the coast of South Wales. Sully Island is accessible by foot when tides are low. But, at high tide, the sea isolates it from the mainland, making crossing by foot impossible. The tourists had prepared for their walk by asking ChatGPT for local tide times. However, when they attempted to return to the mainland, they found the tide rushing in much quicker than they had expected. Locals called a lifeboat to come and rescue the pair. The false information they received from ChatGPT had almost cost them their lives.
This is a concrete example of the danger of using large-language models (LLMs) instead of search engines, independent research or local knowledge. But opting for LLMs also brings a host of other problems, which, although not immediately life-threatening for users, still have a huge impact on people and the environment.
GPTs require more than ten times the energy of a regular internet search. This energy consumption fuels the need for massive data centres. These data centres emit vast quantities of carbon and drain local citizens’ electricity and water resources.
Sticking to regular search engines—crucially, without AI overviews—helps minimise data centres’ impact. But with a range of search engines out there, choosing one that protects your data and meets climate targets, all while returning the information you’re looking for, can be a minefield.
How the AI hype has infiltrated search
As of 2025, Google processes five trillion searches annually. With each search requiring an estimated 0.1-0.2 grams of CO2, the total amount of carbon emitted by finding information on the internet was eye-watering even before AI came along. Now, ever more users choose LLMs over regular search. A 2025 study showed that one-third of respondents would replace search engines with AI tools in the next year. And even with traditional search engines, AI overviews have become a standard feature, making energy-intensive searches even harder to avoid.
The AI hype has led to a global race to stay relevant, with damaging consequences. Fast-tracked data centres from the likes of xAI have sped up operations thanks to smog-emitting generators that harm public health. Meanwhile, in Europe, the EU’s ‘Invest AI’ program has a €200 billion budget “to make Europe an AI continent” by building five AI gigafactories. Non-profit Beyond Fossil Fuels told RESET that “[Big Tech] are actively promoting the integration of generative AI across all sectors of life—public and private, personal and professional—with little regard for whether we can sustain the infrastructure behind this, or whether this type of AI is truly needed across all these areas.”
Adding AI overviews to search is a perfect example of implementing AI features without questioning whether they are helpful or relevant for users. Instead, search engines are motivated by FOMO: a fear of missing out on the profits from users who expect AI overviews when they search. Even so-called green search engine Ecosia has been swept up in the AI hype, adding what they dub ‘the world’s greenest AI’ to their search engine as an automatically active feature. As climate writer Ketan Joshi states, “It really seems like they want the accolades of ‘green AI’ without having to disclose a single actual number [of energy use or emissions].” In its attempts to create an environmentally-friendly AI model, Ecosia appears to have missed the obvious option: the world’s greenest AI would be none at all.
Ecosia and Qwant join forces on a European search index
Relying on US search indexes is a geopolitical risk. Should that access ever be revoked, users outside the US would lose their digital compass, suddenly forced to find information without the tools that define the modern internet. To counteract this ‘walled garden’, search engines Ecosia and Qwant teamed up in 2024 to build a European search index. The result, called the European Search Perspective (EUSP), aims to serve 50 percent of French users and 33 percent of German users by the end of 2025. The EUSP is also selling its cheaper, privacy-focused search index to AI companies, chatbots and search engines that want to offer users accurate search results without relying on Big Tech.
Ecosia and Qwant explain their motivations for developing the EUSP: “Much of Europe’s search, cloud, and AI layers are built on American Big Tech stacks, putting entire sectors—from journalism to climate tech—at the mercy of political or commercial agendas.”
How to choose a search engine
So, how do you cut through the greenwashing and AI hype to choose your go-to search engine? Here are the crucial questions to ask:
- What search index does the search engine use? All search engines answer your queries by calling on an extensive pool of information sources, known as a search index. Many smaller search engines use Google or Bing’s search index to run their own platforms, while others have built their own.
- What’s the environmental impact of the search engine? What data centres power their platform? Transparent disclosure—as opposed to murky greenwashing—counts for a lot here. While some big companies (such as Google or Microsoft) might mention lofty climate targets, consider if they’re taking concrete action right now or simply planning to offset their emissions in the future.
- Does the search engine automatically include AI overviews? While disabling AI overviews is possible (follow these steps for Google search), a platform that uses AI mindfully is a better choice.
- What is the search engine’s privacy policy? Are they selling your data to advertisers? You can always use an ad blocker to stop ads cluttering up your search results, but understanding how a platform intends to use your data is crucial, too.
To find out this information, you need to, ironically, begin with an internet search. Each search engine should disclose its privacy policy and the search index it uses. As for data centres and true environmental impact, that information might be behind closed doors; if so, that’s a red flag. And when a company’s climate policy lists far-off targets rather than the action they’re taking right now, that’s another signal to be wary.
Search engines to check out:
- Mojeek
This UK-based platform doesn’t implement AI overviews, and calls out search engines that do: “Instead of connecting you to the web, these so-called search engines are keeping you away from it [with AI summaries],” it writes on its website. What’s more, it uses its own search index and runs its platform from Custodian, a data centre which uses five times less energy than the industry average. While you will see ads when you search on Mojeek—it uses ads to fund its business model—the platform is firmly against tracking users.
- Brave
You might have heard of Brave, the privacy-focused search engine. But did you know that it’s built its own search index, instead of relying on Google or Bing? While the platform does have AI overviews and shows users ads on the free version, it won’t track your data.
- Qwant
This France-based search engine is in the process of building its own search index in collaboration with Ecosia. In the meantime, it uses Bing to supplement its search results. By default, it doesn’t store your personal data or search history, though you will see AI overviews when you search.
- GOOD
With so many free search engines, it might seem counterintuitive to pay for search. But a paid subscription model, as offered by German search engine GOOD, means you won’t see ads in your search results for as little as two euros a month. The lack of ad trackers also means GOOD can keep CO2 emissions to a minimum.
- MetaGer
MetaGer is another paid model based in Germany, with each search costing one cent. The platform is open-source, non-profit and avoids Big Tech by using Brave and Mojeek’s search indexes.
How can we ensure a green digital future?
Growing e-waste, carbon emissions from AI, data centre water usage—is rampant digitalisation compatible with a healthy planet? Our latest project explores how digital tools and services can be developed with sustainability in mind.
Rethink your relationship with search
It’s not just the environment that’s suffering the negative impact of search engines, but our brains, too. In 2011, a study by Columbia University psychologist Betsy Sparrow proved that humans forget information they believe can easily be found online, dubbing this the ‘Google effect’. Now, with LLMs at our fingertips, we can rely on the internet like never before. With search engines motivated to keep users on their platforms to boost ad profits, it’s up to us to adopt a mindful attitude towards asking the internet for answers.
For the tourists who took a near-fatal walk to Sully Island, help came in the form of a local restaurant owner, who warned them via a megaphone to turn back to the island and wait for a lifeboat to arrive. But ChatGPT’s fabricated tide times serve as a warning against entrusting Big Tech platforms to answer all of our queries. As well as choosing an environmentally and ethically sound search engine, reverting to human knowledge is a powerful way to reduce the digital footprint of search.
