Online publishers are worried that Google AI search will harm website traffic revenue by using content for free

On May 12th, local time, Google announced on Wednesday that it will introduce generative artificial intelligence into search engines, which may be one of the biggest changes in Google's search engine history. But online publishers are generally concerned that the new Google search may affect the traffic on their websites

On May 12th, local time, Google announced on Wednesday that it will introduce generative artificial intelligence into search engines, which may be one of the biggest changes in Google's search engine history. But online publishers are generally concerned that the new Google search may affect the traffic on their websites.

At the annual developer conference, Google announced that it will use artificial intelligence models to integrate and summarize information from various parts of the internet. Google stated that this so-called generative search experience product can better respond to search queries from users.

Google will show some users the text paragraphs generated by AI, and give priority to displaying several relevant links on the search results page, instead of the "ten blue links" usually displayed in Google's search results.

The new Google search based on artificial intelligence is being tested among specific users and has not yet been widely used. But many online publishers have started to worry that if this becomes Google's default search result display method, it will leave more users on Google's website, which may bring less traffic to their website and affect the profits of online publishers.

This controversy also highlights the long-standing tense relationship between Google and its indexing websites, and the emergence of new artificial intelligence tools undoubtedly exacerbates this trend. For a long time, online publishers have been concerned that Google will recombine content fragments on their own websites, but now Google is clearly using advanced machine learning models to "train" artificial intelligence, generating similar text and response results.

Rutledge Daugette, CEO of TechRaptor, a website that focuses on game news and commentary content, said that Google's actions did not take into account the interests of online publishers, and that Google's artificial intelligence search is equivalent to plagiarizing website content.

Their focus is on zero click search, which uses high-quality content created by online publishers and authors with time and effort; unlike users who may click on websites, this does not provide any benefits to online publishers and authors, "said Dougherty, So far, artificial intelligence has been rapidly reusing information from others without bringing any benefits to content owners. In the case of Google, the chat robot Bard may not even provide the source of information used

Luther Lowe, Yelp's director of public policy, has long been critical of Google's search policy. He said that Google's search updates are part of its decades long strategy, aimed at allowing users to stay on Google's website for longer periods of time, rather than directing them to the website that initially provided information.

In an interview, Lowe said, "The exclusivity generated by Google's introduction of the ChatGPT clone into the search field is the final chapter in bleeding the entire network

According to the news website 'Search Engine Land', which closely tracks changes in Google's search engine, AI generated content has so far been displayed in natural search results (valuable free list links related to searches) during testing. It was previously reported that Google planned to redesign its search results page to promote AI generated content.

According to the test of the generative search experience, the AI generated content will be preferentially displayed in the green box at the top of the Google search results page, and there are three boxes on the right showing the links to relevant websites. In the first example of Google search results, the information titles from all three websites were not fully displayed.

Google said that this information was not retrieved from the website, but was only used to confirm the link. SearchEngineLand states that compared to Google's Bard chatbot, the generative search experience is an improvement and a "healthier" way of linking, as Bard rarely directly links to online publishers' websites.

Some online publishers want to know if they can prevent artificial intelligence companies such as Google from crawling content on their websites to train artificial intelligence models. Artificial intelligence companies such as StableDiffusion have already faced lawsuits from data owners, but there is still no clear conclusion on how to define the behavior of artificial intelligence in capturing network data. Reddit and other companies have announced plans to charge for accessing their data.

IAC has multiple websites under its umbrella, including AllRecipe, PeopleMazine, and DailyBeast. Barry Diller, the chairman of the company, is a leading figure in the publishing industry. He said at a meeting last month, "If all the information in the world could be sucked into this big pot and repackaged into declarative statements in the so-called chat function, you could have as much as you want, then there wouldn't be a publishing industry because it's impossible

Diller continued, "What you need to do is gain recognition from the entire industry. Unless you can come up with a system that allows online publishers to obtain payment channels, you cannot steal our content." He said Google will face this problem.

Diller stated that he believes online publishers can sue artificial intelligence companies under copyright law, and currently needs to redefine the scope of "fair use" restrictions. On Wednesday, it was reported that a group of online publisher executives led by Diller said, "If necessary, we will modify the copyright law

The main challenge faced by online publishers is how to determine that the content on their websites is being used by artificial intelligence. Google did not disclose the training source of the large language model PaLM2 behind the generative search experience. Doggett said that although he saw examples of content being rewritten from other websites without attribution on the chat robot Bard, it is difficult to determine whether the information comes from a specific website without a direct link to the source.

Google did not post any relevant comments. PaLM2 is trained based on a large amount of publicly available data on the internet, and we clearly value the health of the online ecosystem. "Google Research Vice President Zoubin Ghahramani said at a media briefing earlier this week," Ensuring a healthy ecosystem is indeed a part of our thinking on how to develop products, and creators are also part of this thriving ecosystem

Doggett stated that Google's move will put independent online publishers in a difficult situation.

I think when many colleagues are fired, we have to worry about our hard work being plagiarized, which is really frustrating for our industry