OpenAI is on the brink of unveiling a search feature for ChatGPT, putting the company more directly in competition with Google and AI startup Perplexity. That framing matters. Search is not merely another checkbox for a chatbot: it is one of the internet’s most valuable and entrenched habits, shaped by decades of users typing a question, scanning links and deciding which sources deserve their trust.
ChatGPT has already changed expectations around how people ask for information. Instead of constructing a search query and opening several pages, users can phrase a question conversationally and receive a synthesized response. The missing piece has been dependable, visible access to current web material and clear sourcing. A ChatGPT search feature that can search the web and cite sources within results would address that gap, while raising the standard by which AI answers are judged.
Table of Contents
OpenAI Ventures into Web Search
OpenAI is developing a feature for ChatGPT that will enable users to search the web and cite sources within search results. The distinction between an answer and a sourced answer is central here. AI systems are useful because they can summarize, explain and reframe information quickly. But those strengths can become liabilities if a reader cannot see where a claim originated, assess the source’s credibility or check whether the material supports the wording of the answer.
The proposed feature would give ChatGPT users access to information drawn from sources such as Wikipedia entries and blog posts. That could make the product feel less like a closed conversational system and more like an interface for navigating the wider web. Yet a source list alone is not a cure for poor search. The meaningful test will be whether citations are relevant, whether they provide enough context for users to verify a response, and whether the system distinguishes between a credible source and a merely available one.
This is where OpenAI’s move becomes more consequential than the language of “AI search” can suggest. Conventional search generally gives users a set of destinations. A chatbot-led search experience attempts to do part of the reading and interpretation before the user arrives. For simple questions, that can save time. For contested, technical or high-stakes subjects, it also makes the quality of the sourcing and the boundaries of the answer far more important.
Multimedia Could Change the Shape of an Answer
The feature is expected to incorporate multimedia elements, including images, alongside textual responses. That is a practical addition rather than a cosmetic one. Some questions are poorly served by paragraphs alone. Someone trying to change a doorknob, for example, may need instructions in sequence, visual references for the relevant parts and diagrams that reduce the chance of misunderstanding a step.
OpenAI’s example of guidance on changing a doorknob points to the broader ambition: to make a response usable, not simply informative. A conventional results page can offer manuals, videos, retailers and forum discussions, leaving the user to sort through them. An AI-led result could instead organize instructions and illustrative diagrams around the task itself.
There is a trade-off. The more complete an answer becomes, the easier it is for users to stop clicking through to the pages that produced the information. That may be convenient for users, but it changes the relationship between search platforms and publishers. Search has long been a referral mechanism as well as an answer engine. If AI systems increasingly summarize source material inside their own interfaces, publishers will be watching closely to see whether attribution translates into meaningful traffic or becomes a small link beneath a finished answer.
Why Perplexity Is the Obvious Comparison
OpenAI’s move comes amid escalating competition in AI-driven chatbots and search engines. Perplexity has surged in popularity and has a staggering $1 billion valuation, driven by its emphasis on accuracy and citation-backed search results. That focus explains why it is a natural benchmark for any OpenAI search product. Perplexity’s appeal is not simply that it provides AI-written responses; it is that it treats citations as part of the experience rather than an afterthought.
For OpenAI, adding search and sourcing to ChatGPT would bring its flagship product closer to the expectations set by search-first AI services. It would also put pressure on rivals. Once users become accustomed to asking a question in natural language, receiving a direct response and checking its sources in the same place, the old divide between “chatbot” and “search engine” begins to look less durable.
Still, competition here is not only about who can produce the smoothest answer. Accuracy, source quality and the handling of uncertainty will determine whether people trust these tools for more than casual questions. Occasional discrepancies in search results, already identified as a challenge, are a reminder that a polished interface does not eliminate the underlying difficulty of retrieving and interpreting information from the web.
Google Is Defending Its Core Territory
Google is not resting on its laurels. The company is working to revamp its core search experience using AI-driven Gemini models, with its latest plans expected at its annual I/O event. Google’s position makes this contest especially charged: search is central to how the company is understood by users, publishers and the wider technology industry.
Google also represents the established model OpenAI and Perplexity are trying to disrupt or reshape. Its familiar results page has trained people to compare sources themselves, even if many users only look at the first few options. An AI response can reduce that friction, but it can also obscure the selection process. The best outcome for users would not be an answer that asks for blind faith; it would be one that is clear about where its information came from and makes verification easy.
The impending I/O unveiling highlights the pace of the contest. Google is bringing Gemini models into its search rethink while OpenAI is preparing its own search feature for ChatGPT. Neither company is treating AI as a side project. Both are responding to a shift in which users increasingly expect search to understand intent, synthesize information and present help in a more conversational form.
A Dormant Page Fuels Speculation
Speculation about OpenAI’s search work has reached a fever pitch on social media, centered on the dormant web page search.chatgpt.com. Observers have noted fleeting redirects to the official ChatGPT website, a detail that has fueled anticipation for an imminent launch. OpenAI has remained silent on the matter.
That silence leaves room for interpretation, and it is sensible to separate observed behavior from confirmed product details. A dormant page and brief redirects can generate excitement, but they do not answer the harder questions about how the feature will work in practice. What matters is whether the eventual product delivers useful search results, relevant citations and a reliable way to move from a generated answer to the underlying material.
The Real Challenge Is Trust
The introduction of a search feature would expand OpenAI’s offerings by giving users a more complete route to online information inside ChatGPT. But “complete” should not mean unquestioned. Search engines and chatbots alike can surface weak material, misunderstand a query or present a confident response that deserves more caution than it receives.
That is why continuous refinement matters. Discrepancies in search results are not a minor technical wrinkle when the product is designed to guide people through online information. Users need enough transparency to recognize when an answer is solid, when the available sources disagree and when they should read further rather than accept a summary.
OpenAI’s planned foray into web search is a pivotal moment in the competition among AI-driven search engines. It is not guaranteed to displace Google, and it does not make Perplexity irrelevant. What it does do is bring a major chatbot platform into a category where answers must be judged not only by fluency, but by their connection to the web and the evidence users can inspect for themselves.
More News: Artificial Intelligence
