OpenAI will not currently expand its user base by releasing AI search engines, as some AI search engines may have limited impact on the market landscape after their release.
On May 10th local time, OpenAI clarified rumors that it would launch an AI search engine.
OpenAI announced on social media that updates to ChatGPT and GPT-4 will be announced live next Monday. “It’s not GPT-5, it’s not search engines. We’re working hard to develop things that people think people will like,” said Sam Altman, CEO of OpenAI.
Previously, there were frequent rumors that OpenAI would launch AI search engines, and suspected grayscale test links were leaked. Multiple sources have reported that OpenAI will release AI search engine products this week at 10 a.m. local time on May 9th. Subsequently, there were reports that the release date would be changed to next Monday, and the product would compete with Google and AI search startup Perplexity, but the release date may change.
OpenAI’s clarification this time may indicate that it still has concerns about the launch of AI search engine products. Previously, OpenAI had made efforts to increase the customer acquisition of large models. On April 1st, OpenAI announced that users could use the ChatGPT generative AI chatbot feature for free without registering, which was seen as a potential challenge to the search landscape. However, currently it appears that OpenAI will not continue to expand its user base by releasing AI search engines.
After the news of not launching an AI search engine spread, Alphabet, the parent company of search engine manufacturer Google, saw a significant increase in its stock price during intraday trading on May 10th, with a narrower decline of 0.77%. OpenAI shareholder Microsoft’s stock price rose slightly by 0.59% on the same day.
The industry has been paying close attention to the news that OpenAI will launch an AI search engine, but at the same time, there are also voices of doubt.
“The AI search engines we see now basically add large model call results and display on the existing search functions. This way of combining search with large models does not change the nature of the search model, and cannot subvert or replace the existing search model.” An AI researcher from an Internet giant told the First Finance and Economics reporter that in a broad sense, the current AI search engine is RAG (search enhancement technology).
RAG is equivalent to helping large models connect to external knowledge bases, guiding the generation of large models by feeding the searched information to them, in order to compensate for the problem of outdated pre training data and inability to provide real-time feedback. The once popular AI search product, Perplexity, uses RAG technology and is built on GPT and Microsoft’s Bing search engine. The above researchers told reporters that the Perplexity solution is not difficult, and the core technology does not have a deep moat.
At the beginning of this year, Jia Yangqing, former Vice President of Alibaba Technology, completed the prototype of an AI search engine in just 500 lines of code for Lepton Search, replicating the solution of Simplicity, sparking discussions in the industry about the AI search engine technology moat.
Currently, some AI search engines have similar functions. The reporter saw on Bing that when searching for page results, the Copilot’s answer and its reference page link were also displayed. Perplexity also generates answers and citation sources after questioning, which can be viewed by clicking on the source link.
“We need to realize that search users are not just about organizing content. Users have many purposes, and some searches are just addressing, which AI search cannot meet. There are also some search purposes that require simple content, and now AI search focuses on analyzing comprehensive results, such as investigation reports.” Fu Sheng, Chairman and CEO of Cheetah Mobile, said in a conversation on May 10th.
In fact, after integrating ChatGPT with Microsoft Edge and Bing search engines in February last year, the impact on the search engine market was limited. According to data from the traffic monitoring platform StatCounter, Bing’s global market share in January last year was 3.03%, which dropped to 2.76% in April and gradually rebounded, reaching 3.64% in April this year. Google’s market share has declined in the past year, but its market position remains stable, with its global search engine market share dropping from 92.82% in April last year to 90.91% in April this year.
Google not only explores AI search engines, but also uses large models to assist advertising and improve ad conversion rates. At present, it is still uncertain whether AI search engines can change the search pattern, and Google has made more significant progress in using large models to assist its advertising business. In the first quarter of this year, Google’s revenue exceeded market expectations, with its advertising revenue increasing from $54.548 billion last year to $61.659 billion. Google’s search and other revenue increased from $40.359 billion in the same period last year to $46.156 billion, a year-on-year increase of 14%.
Not only does OpenAI have some doubts about launching AI search engine services, but Google may also have doubts. There are reports that Google is considering adding some AI search features to its advanced subscription services, and developers are developing related technologies, but the company has not yet made a final decision on when to launch the related services.