As the race for supremacy in the AI-aided search market heats up — accelerated by Microsoft's inclusion of ChatGPT into its Bing search service — Google is ramping up plans to stay competitive.
Company CEO Sundar Pichai revealed last week in an interview with the Wall Street Journal that Google will be adding an AI-driven chat feature to its search service and dismissed notions that chatbots could potentially undermine Google's standing in the search market.
"The opportunity space, if anything, is bigger than before," he told the newspaper. "Will people be able to ask questions to Google and engage with LLMs [large language models] in the context of search? Absolutely."
Pichai revealed that Google is currently experimenting with a range of new search offerings, including versions that permit users to pose follow-up questions related to their initial inquiries.
Although the search giant had previously announced its intention to implement LLMs into search results in a more meaningful manner, it had not yet divulged any information regarding the addition of conversational functionalities.
Google's rushed February announcement of ChatGPT competitor Bard was widely panned after the chatbot displayed incorrect information in its launch ad, underscoring not only its own weaknesses but also fueling additional concerns over the use of chatbots and AI-aided search in general.
Search Grows More Sophisticated with AI
Search has been around as a sophisticated and proven technology for more than 20 years, said Sreekar Krishna, principal and U.S. leader of artificial intelligence for KPMG.
"If you look at what's naturally happened with search during that timeframe, queries have moved away from bringing a million pages for the human user to read through and are now able to highlight where among those pages the answer can most likely be found," he said.
Generative AI pushes search capabilities to the next level, enabling the user to find an answer that may not exist in one source.
It allows for summarized results by pulling together component parts of the final answer from multiple pages while also providing a reasonable way for the user to refer back to these sources.
"In effect, generative AI should increase productivity of the human who engages with search tools," Krishna said.
One of the key success metrics of search is how many times a user needs to re-query a search to get to their final answer, according to Krishna.
"We should all be able to relate to this to some degree if you imagine times where you've had to change your search terms to get to a better set of answers," he said.
The increased time and effort spent re-querying often cause humans to give up on their search because they could not feasibly visit all the pages that popped up and then spend time summarizing their own learnings.
"Generative AI can effectively supercharge search due to its ability to look through vast amounts of information very quickly and summarize findings into a comprehensive answer to meet the search query," Krishna explained.
Chatbots in fact are just the cover on top of the underlying technology that can potentially help humans carry out the next step of the search process, which is to use the insights from the search to do whatever else they were doing — e.g., using search results to buy the specific item the user was looking for or writing a summary of the user's search exercise without having to pull content source by source.
"Search has been a very successful business model for a majority of the modern technology institutions — be it searching the web, searching for movies, searching for products, and so on," Krishna said.
AI Search Could Prove Disruptive, Pave Way for Startups
The monetization of the search has been primarily driven through ads, and these ads have served as a revenue stream for some of the world's largest technology organizations.
But with the ways generative AI is already impacting search, startups are coming out of the woodwork that could potentially disrupt incumbents.
Further, generative AI is starting to lower the bar for machines to understand human spoken and written language.
"Previously, we needed a lot of specialists who could translate from human spoken language to machine understandable languages," Krishna said. "Like software engineers who listen to the requirements in English, French, and then write the code in C, Python, or Java."
But now, the same generative AI that is helping summarize search could also do things like translate from English to Python, for example.
"Such innovation means that our search engines could evolve into task engines," Krishna said. "For example, I could potentially ask for a summary of information, new code to run, or new imagery for writing an article, which could all be done through the task engine's output."
About the author
Nathan Eddy is a freelance writer for ITPro Today. He has written for Popular Mechanics, Sales & Marketing Management Magazine, FierceMarkets, and CRN, among others. In 2012 he made his first documentary film, The Absent Column. He currently lives in Berlin.
