openai is belatedly joining the AI search war. In the early morning of July 26th, Beijing time, openai opened the AI search function. Users can apply for internal testing on the searchgpt official website from now on.
As early as early May this year, there were rumors that openai would enter the search engine market, but it was later denied by openai CEO Sam Altman. The official launch of the AI search function means that OpenAI will challenge Google's search dominance.
According to the official introduction, searchgpt aims to enhance the dialogue capabilities of large AI models with real-time information from the network, providing users with fast and timely answers as well as clear and relevant information sources. Users can follow up on the preliminary answers given by AI search. Openai said it plans to integrate this feature directly into the chatgpt product in the future.
searchgpt demonstration screenshot
Judging from the above official introduction, the general functional form of searchgpt is similar to the existing AI search products on the market. As large models are eager to find application scenarios, productivity tool scenarios including search engines have become the first stop for implementation. Internet giants, AI startups, etc. have entered the AI search track one after another.
According to statistics from the data platform aicpb, in China's domestic market, 360ai search, Mita ai search and Kunlun Wanwei's Tiangong ai search top the list of visits. In fact, chat robot products such as kimi also introduce AI search functions.
In foreign markets, Perplexity, a star company in AI search, has attracted much attention. This start-up company, which was founded in August 2022, has long been ranked second in the AI search list in terms of monthly user visits, second only to New Bing owned by Microsoft.
In late April, perplexity completed its third round of financing of US$62.7 million, and the company’s valuation has exceeded US$1 billion. According to a report by Bloomberg in late June, SoftBank Group’s “Vision Fund II” is investing in perplexity at a valuation of US$3 billion, with an investment amount of approximately US$10 million to US$20 million.
Behind the emergence of many AI search products, there are still many tests such as the establishment of commercial models, the accuracy of content generation, and the balance with the interests of online publishers. The
paid subscription or advertising model is a monetization method for AI search manufacturers. Foreign leading AI search platforms such as Perplexity and You rely on C-side payment to make profits, and users need to become members to obtain a complete functional experience. However, perplexity told the media in April this year that it planned to start selling advertising.
Most of China’s domestic ai search engine products are in the stage of driving traffic for free. However, many practitioners in the field told Nandu reporters that AI search may not be able to bypass the advertising model. A person from a domestic AI search manufacturer previously said that most domestic users subconsciously feel that search engines should be free and the subscription route will be more difficult. "We have made an internal judgment on the commercialization of AI search engines, and the advertising model is still the best and smoothest." Although
ai search can provide the source of the cited information, it cannot overcome the hallucination problem of large models for the time being. Google, which launched AI overviews in May this year, has frequently "overturned" in terms of the accuracy of generated content. Examples of cases that have been ridiculed include when "AI Overview" responded to the question "Cheese can't stick to pizza" by suggesting that users "put some glue on the pizza"; and by mistaking former US President Obama for the "Muslim President".
Liz Reid, vice president and director of search at Google, responded on May 30 that the "AI Overview" not only provides text output, but also includes relevant links so that people can explore further. This means that "AI Overview" generally doesn't hallucinate or make things up like other big-model products. When an "ai overview" goes wrong, it's usually due to misunderstanding the question, misunderstanding the nuances of the language on the web, or not having a lot of useful information available.
In order to deal with the hallucination problem of "AI Overview", Google chose to reduce the application scope of AI search, such as not displaying "AI Overview" of hard news topics, "because the freshness and authenticity of these topics are very important." The conflict between
and the copyright interests of online publishers is also a problem that AI search manufacturers have to face head-on.
html In June, "Forbes" magazine publicly criticized Perplexity for plagiarizing a large part of the "Forbes" article and not citing the original article prominently enough, which may lead some readers to mistakenly believe that the content was written by Perplexity. Perplexity, which was under pressure from the outside world, later revealed that it had reached a revenue sharing agreement with premium content publishers. html On July 17, the Japan Press Association issued a statement stating that the AI search functions launched by Google, Microsoft and other companies use news content as information sources without permission. Many answers given by AI searches are similar to original news articles, which constitutes copyright infringement. Yu. Moreover, many users will not visit recommended websites once they are satisfied with the AI-generated answers, which brings obvious disadvantages to news organizations. The association requires AI search service providers to obtain permission from publishing organizations and pay compensation. It also calls on the Japanese government to amend laws as soon as possible to improve copyright laws and other laws. Whenopenai launched searchgpt, it placed special emphasis on cooperation with publishers and creators. Openai said that searchgpt helps users connect with publishers by highlighting citations and links to original content in search sources, and users can quickly get more results from the sidebar with source links.
"We are working with publishers to build this experience and continue to seek their feedback." Openai tried to show a friendly gesture and also launched a way for publishers to manage how content is presented in searchgpt, giving publishers more choices. Chen Chen, a senior analyst at
Analysys, previously told Nandu reporters that the way AI search directly presents content may affect the interests of traditional content websites in the short term, but it will also give rise to new cooperation models and sharing mechanisms. For example, based on the capabilities of multi-modal large models, AI search service providers can develop advanced search subscription services and cooperate with content production agencies to provide customized content recommendations and in-depth services in professional fields.
Written by: Nandu trainee reporter Yang Liu
