How Does AI Compare to a Standard Google Search? The Environmental Impact Will Shock You

When it comes to the world of search engines, Google has long reigned supreme. But now that people have easy access to AI systems like Chat GPT, some people are opting to take a different approach when they're looking for a fast and easy answer to a question. However, experts are saying that these services may be putting a drain on our natural resources while racing to get us the info we need.

AI vs Google searches, what's the environmental impact? Keep reading to see what the pros say these generative systems are doing to the environment every time we hit "search."

German researchers set out to find the answer to this question, and they published their findings in journal Frontiers in Communication. According to the study, AI uses a grouping of numbers that are called "token IDs" for every word in its prompt.

These token IDs are then sent to data centers, where massive computers work to generate responses through a complex series of rapid calculations.

According to CNN, these data centers have to be quite large to house the computers needed to tackle these tasks, with some requiring more space than a football field. However, the acreage isn't the only resource these big data centers consume.

According to the publication, they also rely on a lot of energy as well, typically that comes from natural gas or coal, which the German researchers say amounts to more than what a simple Google search uses on its own.

According to the study, which involved testing 14 different AI systems with a series of the same prompts, the amount of resources required can vary due to the complexity of the question.

For example, those harder questions produced higher levels of carbon dioxide emissions — levels were up to six times higher than those that allowed for more concise answers — and those AI system that were considered "smarter" produced even more.

Researchers said that the large language models (LLMs) with more reasoning capacities actually produced up to 50 times more of those emissions than their counterparts did, even when answering the same questions.

And that doesn't account for the energy that AI is using to seem more human, either. According to the findings, the AI systems also produce extra emissions when they add in the polite language they've been trained to use by saying things like "please" and "thank you."

That being said, the researchers were quick to point out that not all AI systems are built the same, and those that are designed to be task-specific may be more efficient thanks to their smaller size and the context they use to answer prompts.

Additionally, CNN notes that it's been hard to pinpoint the exact impact of AI due to all of the variables noted, and additional ones that include where a user is located when they use the search engine as well.