Google is overhauling search
Company testing the use of artificial intelligence in its dominant product
Google on Wednesday disclosed plans to infuse its dominant search engine with more advanced artificialintelligence technology, a drive that’s in response to one of the biggest threats to its long-established position as the internet’s main gateway.
The gradual shift in how Google’s search engine runs is rolling out three months after Microsoft’s Bing search engine started to tap into technology similar to that which powers the artificially intelligent chatbot ChatGPT, which has created one of Silicon Valley’s biggest buzzes since Apple released the first iPhone 16 years ago.
Google, which is owned by Alphabet Inc., already has been testing its own conversational chatbot called Bard.
That product, powered by technology called generative AI that also fuels ChatGPT, has only been available to people accepted from a wait-list. But Google announced Wednesday that Bard will be available to all comers in more than 180 countries and more languages beyond English.
Bard’s multilingual expansion will begin with Japanese and Korean before adding about 40 more languages.
Now Google is ready to test the AI waters with its search engine, which has been synonymous with finding things on the internet for the past 20 years and serves as the pillar of a digital advertising empire that generated more than $220 billion (U.S.) in revenue last year.
“We are at an exciting inflection point,” Alphabet CEO Sundar Pichai told a packed developers conference in a speech peppered with one AI reference after another. “We are reimagining all our products, including search.”
More AI technology will be coming to Google’s Gmail with a “Help Me Write” option that will produce lengthy replies to emails in seconds, and a tool for photos called “Magic Editor” that will automatically doctor pictures.
The AI transition will begin cautiously with the search engine that serves as Google’s crown jewel.
The deliberate approach reflects the balancing act that Google must negotiate as it tries to remain on the cutting edge while also preserving its reputation for delivering reliable search results — a mantle that could be undercut by artificial intelligence’s penchant for fabricating information that sounds authoritative.
The tendency to produce deceptively convincing answers to questions — a phenomenon euphemistically described as “hallucinations” — has already been cropping up during the early testing of Bard, which like ChatGPT, relies on stillevolving generative AI technology.
Google will take its next AI steps through a newly formed search lab where people in the U.S. can join a wait-list to test how generative AI will be incorporated in search results.
The tests also include the more traditional links to external websites where users can read more extensive information about queried topics. It may take several weeks before Google starts sending invitations to those accepted from the wait-list to test the AI-injected search engine.
The AI results will be clearly tagged as an experimental form of technology and Google is pledging the AI-generated summaries will sound more factual than conversational — a distinct contrast from Bard and ChatGPT, which are programmed to convey more humanlike personas. Google is building in guardrails that will prevent the AI baked into the search engine from responding to sensitive questions about health — such as, “Should I give Tylenol to a 3-year-old?” — and finance matters. In those instances, Google will continue to steer people to authoritative websites.