"Come Back in a Week": Snap a Photo and Instantly Find Out Cherry Blossom Peak Season [AI Assistant Era]③
The Evolved AI Assistant: Google Search Live
Snap a Photo and Ask with Your Voice for Instant Answers
"It is currently the early stage of blooming, with the flower buds just starting to open. Numerous pink buds have formed on the trees, and since the flowers will be in full bloom about a week after this stage, you can expect to see them at their peak if you visit Yeouido between April 7 and 12."
At Yeouido Hangang Park in Yeongdeungpo-gu, Seoul, someone is searching on Google Search Live on the 2nd. It shows the flowering status and full bloom timing of flowers. Photo by Eunseo Lee.
View original imageOn the afternoon of April 2, in front of Yeouido Hangang Park in Yeongdeungpo-gu, Seoul, where cherry blossoms had just begun to bloom, I opened Google. After photographing a cherry tree with the rear camera of my smartphone, I asked how much the flowers had bloomed and when would be the best time to come back for optimal flower viewing. It also promptly answered how much later the cherry blossoms would bloom compared to the nearby Cornelian cherry trees. Rather than simply searching with a phone, it felt as if a smart, invisible assistant was right next to me, providing answers to all my questions.
Google launched SearchLive in Korea on March 27. This search feature enables real-time, interactive conversations by alternating between voice and camera inputs. After its initial introduction last year in the United States, the service has now been expanded to Korea. In 2024, Google released "AI Overview," which summarizes the key information of search results using generative AI, and last year, it supported "AI Mode," which strengthened inference and multimodal capabilities. This time, Google has introduced a service that can recognize the situation like an assistant present in the same location as the user and provide contextual answers.
At the core of this change is the multimodal function, which utilizes unstructured data such as images and voice to enable complex searches beyond text. The key feature of this SearchLive is that "context-based conversation" is now expanded through voice and camera recognition. Unlike in the past, when users had to type keywords into a search bar, they can now receive only the most relevant information in a conversational Q&A format tailored to the situation.
The convenience of being able to ask questions conversationally without typing keywords by hand, even while on the move, is a strong advantage of SearchLive. Users can click on the links provided in the answers to find more in-depth information. It was also possible to ask emotionally nuanced questions, such as "Why do cats walk around so confidently even though they don't pay for anything?" or "Please recommend some hip-hop music that suits today's afternoon weather." In situations that are difficult to describe in words or text, activating the camera allows SearchLive to answer questions based on visual context.
This is a script conversion of the SearchLive results received by voice. It provides a detailed explanation of the origin and taste of Financier. Photo by Eunseo Lee.
View original imageThe quality of searches has also improved. By combining real-time video with voice, users can convey their intended context in greater detail. It felt as though the service provided comprehensive insights, rather than just single pieces of information. For example, when I took a picture of a pecan financier with the camera, it explained, "It's a dessert topped with plenty of pecans, so the sweet flavor is complemented by a nutty texture," and added, "'Financier' is a French word meaning 'financier'; the pastry is named for being baked in a gold bar-shaped mold." In addition, I was able to ask about the origin, growing conditions, and health status of plants at home, or have it find products tailored to my preferences and single-person household situation.
Access to the Latest Information Is Limited... 'Latest' Means Up to 2023–2024
However, not all search results were based on the latest information or accurate research and reporting. Upon checking the cited sources in the answers, most links were from Namuwiki, Facebook, Tistory, etc., and the information provided was mostly from 2023 to 2024. For location-based questions, inaccurate responses sometimes occurred. For example, when I asked, standing near Seoul City Hall in Jung-gu, "How long does it take to walk from here to Gwanghwamun Station?" the starting point was incorrectly set to Gongju in Chungcheongnam-do, which was confusing. When I asked how long it would take to walk to a supermarket about 10 minutes away, the answer was that I could arrive in two to three minutes.
Hot Picks Today
"Two Spend 500,000 Won in a Flash, Yet Wait Two Hours?" ... 'Dreamland' Shunned by Kids and Foreign Tourists
- The Era of Getting 200 Million Won After 30 Years Is Over... A Japanese Bank's "Perpetual Department Head" Who Just Clocked In Cries Out
- Spain Hit by Its Worst Tragedy... Lee Kangin, Moreno and Real Madrid Join Soccer World in Mourning
- They Said, “Make It Your Home for Life” - Then Suddenly Kicked 200 Passengers Off... What Happened to the Around-the-World Cruise?
- "130 Rice Bowls Sold Out in a Day"... 'Unfiltered Face' Adds 8 Million Won to Night Market Sales: China's Bungeunwoo
When restoring the Google Search Live results into a script, part of it does not appear, and the message "An error occurred, AI cannot generate a response" is displayed. Photo by Eunseo Lee.
View original imageIt was also difficult to check voice-provided search results in script form. After the conversation ended, when I wanted to review the content as text, some parts were inaccessible, displaying the message, "An error occurred, AI cannot generate a response."
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.