The internet search wars are back.
The emergence of AI systems capable of generating direct textual answers to questions (most notably the ChatGPT chatbot created by San Francisco-based OpenAI) has prompted Google to take on a collaborative challenge from Microsoft’s Bing. It opened the first new front in the battle for search supremacy since dodging. ten years ago.
Google and microsoft Both are about to announce a revamp of their search engines to include direct answers provided by artificial intelligence. Meanwhile, some search startups have already incorporated AI into their services, technology behind ChatGPT has the potential to transform one of the largest online marketplaces.
Analyst Greg Sterling, who has followed the search market since 1999, said the sudden surge in experimentation is a long one. “People are ready for something simpler, deceptively more reliable, and less ad-laden.”
Systems such as ChatGPT, which are based on so-called large-scale language models that can “understand” complex queries and generate text responses, are not direct replacements for search by themselves. The information used to train ChatGPT is at least a year old, and the answers ChatGPT provides are already “in memory” rather than more targeted material pulled from the web in response to a specific query limited to information.
This has led to a race to develop new hybrids of AI and traditional search. This technique, known as search augmentation generation, first applies a search tool to identify the pages containing the most relevant material, then uses natural language processing to “read” them. The results are injected into a large language model, such as his GPT-3 in OpenAI, to spit out more accurate answers.
Google’s rivals say this has led to a rare opportunity to create a new and distinctive search service, rather than trying to beat the search giant at its own game.
Sridhar Ramaswamy, a former Google executive and co-founder of search startup Neva, said, “Google used to try to beat them with a better set of links than anyone, including us. So I got over it.
His company has started inserting short text answers into search results pages. To overcome the weaknesses inherent in large language models (which can produce erroneous results), Neeva first added citations to text answers to indicate the sources behind the claims. rice field.
On the other hand, some companies have taken a more radical approach by giving users a pure chat interface similar to ChatGPT and eliminating the search results page altogether.
Microsoft plans to allow users to “switch” between the chat interface and the traditional search results page. Revamped Bing service Late last week. The software company has refused to verify that screenshots of apparent online leaks are genuine. Added a “chat” button to toggle.
Pure chat services are far more appealing to younger generations looking for a completely different experience, said Angela Hoover, co-founder of Andi, another search startup.
“Gen Z is hungry for something new, so they’re turning to TikTok,” she said. I would like an answer.”
Aravind Srinivas, co-founder of Perplexity AI, says the full commitment to conversational interfaces will also lead to more distinctive and useful services. “It’s basically Wikipedia on steroids, transcending the concept of search engines,” he said.
However, the benefits for companies using AI to supplement search results may be short-lived. Google said last week that it would “soon” start including the results of its own language AI system in its results, allowing users to interact directly with the new AI technology.
That could make it difficult for competitors to come up with something distinctive enough to break Google’s dominance in the market, Sterling said. “If Bing integrated AI into search the way Google did, we might not see any change,” he added.
But even if the impact on competition proves to be limited, the broader economic impact on the web still looks grave. If so, there is a risk that internet users will find the answer without clicking through to the page below.
Perplexity’s Srinivas said: “This is not good for publishers.”
He predicted that this would force publishers to rethink their relationship with search engines, either shutting down their sites or trying to negotiate fees for allowing information to be crawled and processed.
For search engines, on the other hand, the higher cost of generating full-text answers can jeopardize what was once a highly profitable business. Srinivas estimates that AI-powered searches could cost him seven to eight times more than traditional internet searches.
“These will only succeed if the value per query is much higher,” he said.
For Microsoft, which spent billions of dollars competing with Google in the early days of Bing, lower profit margins may be an acceptable cost to its chances of ultimately gaining significant market share. If Microsoft has helped undermine a key pillar of its rival’s business, lowering profit margins for everyone could serve a broader strategic purpose as well.
Meanwhile, most search startups say they want to avoid relying on advertising just to differentiate themselves from Google.
Andi plans to use a “freemium” business model, charging some users for higher levels of service, Hoover said. “We never want to run ads, which she thinks is the one that destroyed Google search,” she said.
But none of the companies experimenting with AI in search have yet tested whether consumers are willing to pay for what they currently receive for free. According to Srinivas, Google’s founder himself once wrote that relying on advertising could be the wrong business model for search engines. At this stage, he added, it’s better to consider all options.
