Google AI Search Glitch Turns Queries Into Chatbot Conversations with Certain Keywords

A peculiar bug has been discovered in Google’s AI search function, “AI Overviews.” When users search with certain English keywords, instead of displaying a summarized search result, the system responds as if they were conversing with an AI chatbot.

What Happens When You Search for “disregard”?

On Friday, May 22, 2026, this issue was highlighted on social media and subsequently investigated by technology news outlet The Verge. Typically, when a user enters the word “disregard” into Google Search, the AI Overviews feature should display a summary of the word’s definition and related details. However, The Verge found that the following message appeared instead: “Understood. Let me know if there’s anything else you need or if you have a new question later!” This response clearly mimics the kind of interaction one would expect from a chatbot interface and is unrelated to the definition, examples, or links to related content that the user was likely seeking.

Similar Issues with “ignore” and “skip”

The bug is not limited to the word “disregard.” When The Verge’s reporters tested the word “ignore,” the AI Overviews feature displayed the following message: “Message received! I’m ready to help. What would you like to focus on today? Let me know if there’s a specific topic, task, or question!” Similarly, when the word “skip” was entered, the system produced a response along these lines: “This message seems to be a test or a typo! Let me know if there’s a question or prompt you’d like to share or if there’s anything I can assist with. I’m always ready!” In all these cases, the output deviates significantly from what users would expect from a search engine’s AI summarization functionality, instead resembling a chatbot’s conversational response.

Google’s Response and Current Status

After the issue gained widespread attention, Google took action by removing the AI Overviews feature from the search results for the term “disregard.” As of the article’s publication (May 22, afternoon), searching for “disregard” brought up a list of news articles about the bug rather than AI-generated summaries. However, for “ignore” and “skip,” the anomalous chatbot-like responses in AI Overviews persisted at the time of publication. The Verge reached out to Google for comment but had not received a response by the time the article was published.

Why Did This Bug Occur?

Google’s AI Overviews is designed to collect, summarize, and present web-based information in response to search queries. Since its full-scale launch in the U.S. in 2024, the feature has faced occasional criticisms regarding its quality. In this case, the bug likely stems from the system interpreting the search queries as “commands” rather than keywords. Words like “disregard” (“to ignore”), “ignore” (“to pay no attention to”), and “skip” (“to omit”) may have been processed by the AI not as search terms but as directives, inadvertently triggering the chatbot response mode. It is assumed that the large language model (LLM) powering AI Overviews operates in different modes for summarizing search results and engaging in conversational responses. Specific keywords may have disrupted the logic that determines which mode to activate, resulting in this error.

The Reliability of Google’s AI Search

Although this bug is relatively minor and likely to be fixed quickly, it underscores the structural challenges facing Google’s AI search functionality. Since its comprehensive rollout during the 2024 Google I/O conference, AI Overviews has become a central part of Google’s search experience. However, it has faced significant criticism from the outset, with bizarre incidents such as generating recommendations like “eating rocks” or “spreading glue on pizza.” While Google has made ongoing efforts to improve the quality of AI Overviews, this bug highlights a fundamental limitation of LLMs: they cannot always interpret inputs in the correct context.

Impact on Users

For regular users, encountering this bug could undoubtedly cause confusion. If someone searches for the meaning of a word, only to receive a response like “Understood. Let me know if there’s anything else I can help with,” it offers no assistance. Furthermore, when the AI responds with “Is there anything else I can help you with?” it disrupts the user’s expectations, as they are using a search engine, not interacting with a chatbot. More concerning is the potential erosion of trust in AI Overviews. Many users perceive AI-generated summaries as “Google-endorsed information.” If the AI fails to correctly interpret basic queries, confidence in the entire system may be undermined.

Future Outlook

The speed with which Google can fix this bug and implement measures to prevent similar issues in the future will be closely watched. As LLM-based search functions continue to evolve, managing the boundary between “search” and “conversation” remains a key technical challenge, not just for Google but for all platforms developing AI search technologies. A reporter from The Verge noted, “This is an amusing incident, but it’s likely just a bug that Google will fix soon.” While this may be true in the short term, Google must remain vigilant to ensure that such “amusing bugs” do not accumulate and lead to a catastrophic loss of user trust.