Changing Search Trends in the AI Era: Insights from Data

"Recently, I feel like I'm searching less frequently" — many people probably share this exact sentiment.
With the emergence of generative AI, an increasing number of people say they now turn to AI chatbots for quick research.

In fact, global search usage is beginning to show signs of fluctuation.
A survey by Semrush reported that Google's global traffic decreased by about 8% after the release of ChatGPT. On the other hand, an analysis by SparkToro confirmed that the actual number of Google search queries increased by 21% from 2023 to 2024.

In other words, we are currently seeing a seemingly contradictory trend: while the number of "people accessing search" is decreasing, the "number of search queries issued" is actually increasing.

So, what kind of changes are occurring in Japan at the individual keyword level?
By comparing recent search data with the past, we would like to dive deeper into the following three points.

1. Trends in general proper nouns and system-related keywords
2. The decline of "recommendation" type keywords
3. The rapid surge of AI-related keywords

1. Search Trends for General Keywords

When examining the Top 1,000 keywords with likely high search volumes generated by GPT, keywords directly linked to daily life and official systems, such as "Aeon," "Library," "Pension," and "Don Quijote," showed a stable or increasing trend.

For "themes requiring accuracy and the latest information," such as business hours and institutional details, search engines are more suitable than AI, which is why search is still considered to be widely used. Of course, some keywords are decreasing, but this is thought to be driven more by seasonality and market trends rather than AI substitution.

2. Search Trends for "Recommendation" Type Keywords

Next are comprehensive keywords such as "recommendations," "rankings," and "comparisons."

These have shown a clear year-over-year decline, with recent data revealing many cases of double-digit percentage drops compared to the past.

Generative AI is the driving force behind this.
If you directly ask an AI, "What are the recommended XX?", a summarized answer is returned instantly, reducing the need to intentionally search and look for comparison articles.

3. The Rapid Surge of AI-Related Keywords

And the category with the most momentum is AI-related words.

Terms like "ChatGPT," "Generative AI," "Stable Diffusion," "Claude," and "Gemini" had virtually zero search volume until 2022, but surged rapidly from 2023 onwards. Recent data indicates they have reached scales of millions to tens of millions per month.

Additionally, peripheral words such as "free," "app," and "how to use" are growing alongside them, suggesting that AI is becoming firmly established as a part of daily life and work.

4. Summary

Comparing recent search data with the past reveals that a restructuring of search behavior is underway.

• General proper nouns and institutional words → Stable or increasing
• "Recommendation" type words → Significant decline
• AI-related words → Explosive growth

Even globally, there is a mix of data suggesting that search usage itself is "decreasing" and data suggesting it is "actually increasing," which means we can safely say that the role of search is currently in the midst of a transformation.