TOKYO - AI is beginning to transform how people shop online, with a growing number of users turning to tools like ChatGPT instead of traditional search engines such as Google. According to e-commerce consultant Tomoyuki Mochizuki, this trend is especially noticeable among younger consumers in Japan.
In a recent survey of Japanese men and women in their 20s to 60s, conducted by Mochizuki’s firm, over half said they use platforms like Rakuten and Amazon to search for products. While Google remains strong, around 7% now use AI-based search, with the proportion rising to 20% among those in their 20s.
Mochizuki noted that AI searches are particularly useful for more complex or subjective needs, such as finding a "stylish portable power station"—a phrase difficult to interpret through conventional search. AI tools interpret such requests more intuitively, offering results based on implied aesthetics like wood finishes or dark tones.
Data also show a sharp rise in e-commerce traffic from AI searches. Over the past year, traffic to Amazon via AI searches has increased 13-fold, while Rakuten has seen a 69-fold surge. The boost began around October last year, possibly linked to the launch of new AI models and broader corporate adoption.
OpenAI recently announced that ChatGPT would support integrated shopping capabilities, allowing users to view product images, prices, and reviews within the chat window. In the U.S., a new feature even enables purchases directly within ChatGPT. While this function is not yet live in Japan, it could signal a broader transformation of AI as a commerce platform.
Security concerns remain, particularly around storing personal data like credit card information. However, current implementations typically rely on external platforms such as Amazon for payment processing, avoiding storage of sensitive information within ChatGPT itself. Users are encouraged to teach the AI about their preferences through usage history while avoiding the input of private details.
Industry observers believe OpenAI aims to evolve from a pure chat tool into a comprehensive platform, potentially earning revenue through affiliate commissions or advertising in the future. Despite expectations, no ads have yet appeared in ChatGPT’s interface.
Google, too, is adapting. Since August last year, it has rolled out an "AI Overview" feature that summarizes search results at the top of the page. While not yet widespread for shopping queries—possibly to protect advertising revenue—it already affects user behavior. Research shows that up to 30-40% of users skip the traditional link list below the AI-generated summary.
Experts caution that the AI Overview can sometimes be inaccurate. Mochizuki cited a personal case in which Google incorrectly displayed the Japanese calendar year, emphasizing the need to verify facts independently, especially in academic settings.
The opaque nature of AI systems means that even researchers often struggle to understand how results are ranked or generated. Claims by firms that they can optimize content for AI visibility are viewed with skepticism, as the algorithms behind these features are not disclosed.
While the rise of AI search is likely to continue, its utility may be limited in domains where physical inspection is vital, such as book purchases or color-sensitive cosmetics. In these areas, users still prefer visiting stores or watching product videos on platforms like YouTube and TikTok.
Looking ahead, AI search is expected to become increasingly specialized, with different tools serving distinct product categories and user habits. Mochizuki advises consumers to use AI tools extensively to personalize their experience—while remaining cautious about data privacy.
Source: ABEMA
