OpenAI discloses ChatGPT has reached 1.2 billion weekly active users, with search query volume about 20% of Google's

The data was disclosed during an OpenAI event held in New York on November 2. Adam Fry said ChatGPT's growth rate is accelerating further. According to him, ChatGPT's weekly active users grew from 1 billion to 1.2 billion over the past 60 days, equivalent to adding about 100 million users per month.

One of the most noteworthy figures is the growth of ChatGPT's search business. Fry revealed that OpenAI tallied search-type queries within the platform and compared them with Google, finding that ChatGPT currently handles search-type query volume equal to about 20% of Google's.

He explained that "search" here is not limited to keyword retrieval in the traditional sense, but refers to situations where, after a user asks a question, ChatGPT needs to call on web search to obtain the latest information. Whenever a user asks a question that requires an internet search, it is counted as one search.

This shift also reflects a change in how users use ChatGPT. Initially, users mainly used it for writing assistance and programming, tasks that are not typical use cases for traditional search engines. But over the past year, more and more users have begun using ChatGPT to look up people, places, products and other content requiring real-time information, and questions in the finance and health domains have gradually become important use cases.

Fry believes that one of the biggest differences between ChatGPT and traditional search engines is its ability to retain a user's long-term context. For example, when a user asks about the height of the Eiffel Tower in Paris, the system can not only answer with the specific number but also combine it with the Paris travel plans the user mentioned earlier to provide information that better fits actual needs. This ability to continuously understand a user's background and goals is gradually transforming ChatGPT from a one-off question-and-answer tool into an assistant capable of participating in long-term tasks.

Voice features have also achieved significant growth. Currently, ChatGPT's combined weekly active users for voice input and voice conversation reach 250 million, about one-fifth of overall weekly active users. Fry said voice is no longer just a novelty feature users occasionally try, but is gradually becoming an everyday way of using the product. Some users simply ask a quick question by voice, while others hold longer conversations.

He attributed this growth to the improved interaction experience brought by the new generation of voice models, especially more natural two-way communication. The system can respond appropriately during a conversation and can also handle users interrupting and adding information. In addition, background voice mode allows users to continue a conversation even after closing the app or moving around the house.

On the personal intelligent assistant front, Fry also discussed Dots, which OpenAI launched on September 29. Unlike traditional chatbots, Dots is positioned as a personal assistant that can run continuously and proactively offer help. For example, users can ask it to remind them at the end of each day whether they have missed anything important, and the system can combine information from emails and messages to help users review the day's to-do items.

However, OpenAI is still quite cautious about permission management for Dots. Fry said a personal assistant needs to handle meaningful matters on behalf of users, so users must build sufficient trust in it. Compared with simply answering questions, having AI perform operations on a user's behalf involves a higher degree of responsibility and risk.

He also acknowledged that Dots is currently a product with relatively high computing resource consumption. OpenAI wants to first ensure it can provide a sufficiently good service, then gradually optimize operating costs. As the new generation of models continues to improve efficiency, it is expected that computing demand can be reduced in the future while maintaining service quality.

Regarding Dots' business model, Fry said OpenAI is building revenue sources through multiple means such as subscriptions, advertising and additional usage credits. He revealed that ChatGPT's previously announced paid subscribers had exceeded 50 million, and that this figure has since grown quite significantly. The advertising business achieved more than $1 billion in annual recurring revenue (ARR) within 200 days of launch.

On product strategy, Fry also confirmed that OpenAI plans to integrate ChatGPT's current Chat and Work modes into a unified interface.

He acknowledged that requiring users to choose between Chat and Work has become a point of confusion in using the product. Users are often unclear about when they should use ordinary chat and when they should switch to Work mode, which can execute complex tasks.

OpenAI hopes to have the model itself judge user intent, automatically distinguishing simple Q&A from complex tasks requiring long execution, thereby eliminating this mode switching. Fry said this adjustment is underway and is expected to launch soon.

Image generation is also an important growth area for ChatGPT. Currently, users generate 1.5 billion images per week through ChatGPT. Fry said that initially users mainly used image generation for entertainment, such as making retro-style personal photos and family posters, but today the scope of application has expanded to work scenarios, including making restaurant menus, promotional posters, birthday posters and various visual charts.

OpenAI is also continuing to reduce image generation response times, hoping to further expand use cases by improving speed. Regarding user feedback that AI-generated restaurant menus and signs look too similar in style, Fry agreed and promised to pass the relevant feedback to the product team.

He is particularly optimistic about the direction of interactive interfaces. Fry believes that over the past few years AI chat products have mainly relied on text for interaction, but in the future they will gradually shift toward richer visual experiences. For example, when a user asks by voice about the offside rule, the system should be able to immediately generate a diagram; if the user still cannot understand, it can readjust the diagram based on subsequent feedback.

He believes that this kind of interactive content, which can continuously update as a conversation unfolds, will play an important role in education and knowledge learning.

Finally, when it comes to public concerns about artificial intelligence, Fry acknowledged that AI is changing the way work is organized, and that this change is indeed unsettling for many people. But he believes that as more people use AI themselves and understand the help it can provide in daily life, public understanding of this technology will gradually change.

Judging from these figures and product plans, OpenAI is pushing ChatGPT to transform from a traditional chatbot into a comprehensive personal assistant. Search, voice, image generation and automated tasks are gradually converging, and in the future users may no longer need to distinguish whether they are asking a question, searching for information or asking AI to carry out a complex task; they will simply express their ultimate goal, and the system will decide how to accomplish it.