Google's New AI Search Feature Is Breaking Dictionary Lookups. Here's Why That Matters

Google's AI Overviews feature—a new tool that automatically summarizes search results for you—is breaking something simple: looking up word definitions. When you search for a word like "disregard," Google's AI now treats it as an instruction to ignore something instead of finding the word's definition. For over 20 years, Google has reliably shown dictionary definitions when you searched for a single word. That is no longer guaranteed.

9to5Google documented the problem. The AI Overviews feature, which launched widely across Google Search, is interpreting your search in a conversational way—like you are talking to a chatbot—rather than matching your exact words to dictionary entries. Previously, Google would show you a dedicated dictionary box with definitions from sources like Merriam-Webster.

This is a sign of how Google's search engine is fundamentally changing. The company's AI systems now try to guess what you mean rather than simply match the words you typed.

How This Happened

AI Overviews is now a permanent part of Google Search for all U.S. users. Instead of searching the internet for pages that contain your exact words, the feature uses AI to generate a summary of what it thinks you are asking.

Here is the key point: you cannot turn this feature off completely. According to Google's support pages, the feature affects everyone. This is not something you can opt into or out of. The dictionary lookup problem is now a system-wide issue.

Why Dictionary Lookups Broke

The problem comes down to how AI learns to understand language. Words like "disregard" and "ignore" are tricky because they can mean two different things. They are definitions you might look up in a dictionary. But they are also instructions you might give to a chatbot—like saying "ignore what I just said" in a conversation.

Google's AI system is now trained to prioritize the conversational interpretation. It thinks you are giving it an instruction, not asking for a definition. The system does not have enough context awareness to tell the difference.

What Google Says About Accuracy

Google claims that AI Overviews are as accurate as its old Featured Snippets—those highlighted boxes that used to appear at the top of search results. The company also reports that people spend more time on web pages when they have read a summary first, which suggests the feature is helpful.

The dictionary lookup issue, though, signals that Google's AI still struggles with words or phrases that have multiple meanings. As the AI system evolves, there may be other surprises like this one.

The Bigger Picture

This change reflects a fundamental shift in how Google thinks about search. For 25 years, search has worked like this: you type words, Google finds pages with those words, you click a link. Now Google is trying to understand what you actually want and give you an answer directly, the way you might ask a friend a question.

The broader context here is worth noting. We have seen Google change how search works before—when it moved from simple lists of links to rich information boxes and knowledge panels in the early 2010s. Each time, people had to adjust. The difference now is that you cannot choose to opt out, and the changes are happening faster.

For website owners and content creators, this shift matters too. The old rules for getting your content to rank in Google are no longer the only rules. Your content now needs to work for an AI system that is trying to summarize your information, not just match keywords.

What Happens Next

Google has tools built in to prevent its AI from being tricked or giving harmful advice. But the dictionary lookup problem shows that these safeguards do not catch everything. Some questions and words—the edge cases—still confuse the system.

Google also offers a more experimental AI search mode called AI Mode through its Search Labs program. This version can search the web differently and will show you links when it is not confident in its own answer. But you have to choose to use it separately.

The bottom line: Google is rebuilding search around AI, not traditional keyword matching. That is a big change. It may work better for many kinds of questions. But as the dictionary lookup issue shows, it is still being figured out. Some established search behaviors will break before new ones settle in.