[Research Release] Web Articles in the AI Era: Top Reason for User Abandonment is "Unclear Conclusion" at 44.8% -- "AI-likeness" Ranks 7th (17.1%)
A survey by Thinkmove Inc. reveals that the top reason for user abandonment of web articles is "unclear conclusion" (44.8%), while "AI-likeness" ranks 7th (17.1%). The study also found that readers who frequently use AI tend to be more critical of web articles.
📋 Article Processing Timeline
- 📰 Published: May 8, 2026 at 17:21
- 🔍 Collected: May 8, 2026 at 08:31
- 🤖 AI Analyzed: May 8, 2026 at 08:33 (1 min after Collected)
Source: Thinkmove Inc. "Survey on Web Content Reading Habits in the AI Era" (n=210, multiple responses)
Thinkmove Inc. (Headquarters: Shibuya-ku, Tokyo; Representative Director: Shota Toyokura) conducted a screening survey (n=1,000) targeting men and women aged 15-99 nationwide, and a main survey (n=210) targeting those who read web articles and have experience with generative AI, to understand the actual state of reading web articles after the spread of generative AI.
This survey revealed that the top reason for readers abandoning web articles was "unclear conclusion or answer" (44.8%), while "feels like an article written by AI/robot" accounted for only 17.1%. Furthermore, a cross-tabulation of respondents who answered "significantly increased/slightly increased" their AI usage frequency (n=137) and those who answered "unchanged" (n=53) confirmed that the group with increased AI usage had higher abandonment rates for almost all reasons for abandonment.
## Key Points of This Survey
- The top reasons for abandoning web articles were not AI-likeness (17.1%) but "unclear conclusion" (44.8%), "content duplication" (43.8%), and "strong advertising tone" (38.6%).
- It was confirmed that readers who use AI more frequently tend to be more critical of web articles regarding "information freshness," "content relevance," and "readability" (with a maximum difference of 19.9 percentage points).
- However, for "AI-likeness" alone, the difference between the increased AI usage group and the unchanged group was small at 3.2 percentage points, indicating that sensitivity to "whether it is AI-generated" itself did not change significantly even with increased AI usage.
## Information on the Detailed Report
The results of all 12 questions × all options not listed in this release, cross-sectional analysis of questions, and details of the questionnaire are available on the detailed report page below.
"Detailed Report on Web Content Reading Habits in the AI Era"
https://thinkmove.jp/survey/ai-contents-survey/
## Survey Overview
### Fact 1: The top reason for reader abandonment is not "AI-likeness"
When respondents to this survey (n=210) were asked (multiple responses) why they "abandon (stop reading)" web articles, the highest response was "unclear or vague conclusion/answer" (44.8%). This was followed by "same content repeated multiple times" (43.8%) and "feels like it's for advertising/affiliate purposes" (38.6%).
On the other hand, "AI-likeness," which has been discussed in the content industry in recent years, meaning "feels like an article written by AI/robot," was 17.1%, ranking 7th out of 9 options.
Source: Thinkmove Inc. "Survey on Web Content Reading Habits in the AI Era" (n=210, multiple responses)
All top three abandonment reasons are related to article structure, information design, and advertising expression, not the writing style itself. It can be concluded that readers abandon articles when they find it difficult to reach valuable information, such as "cannot reach the conclusion," "same content is repeated," or "the boundary between advertising and main text is ambiguous."
However, "being AI-generated" does not entirely have no impact on reader evaluation. In another question (Q8) of this survey, responses to knowing an article was AI-generated included "reading priority decreases, often search for another article" (12.4%) and "close the page" (4.3%), indicating that a certain number of readers are wary or abandon content specifically because it is AI-generated.
### Fact 2: Readers who use AI more frequently are more critical of web articles
This survey conducted a cross-tabulation of question Q11, which asked respondents "how their frequency of using AI for information gathering has changed compared to 1-2 years ago," and Q7 (reasons for abandonment).
Comparing the AI usage increase group (n=137) who answered "significantly increased/slightly increased" and the "unchanged" group (n=53), it was confirmed that the group with increased AI usage had higher abandonment rates for almost all reasons for abandonment.
Source: Thinkmove Inc. "Survey on Web Content Reading Habits in the AI Era" (cross-tabulation, multiple responses)
The largest differences were observed in "information is old/not updated" (+19.9 percentage points), "content was different from what I wanted to know" (+12.7 percentage points), and "difficult to read/hard to see" (+9.8 percentage points). Dissatisfaction regarding information freshness, relevance, and readability was significantly higher among the group with increased AI usage.
This result suggests that readers accustomed to instantly getting answers from generative AI may raise their bar for web articles, questioning "is there valuable information worth reading here?"
### Fact 3: Sensitivity to "AI-likeness" does not significantly change with AI usage frequency
Thinkmove Inc. (Headquarters: Shibuya-ku, Tokyo; Representative Director: Shota Toyokura) conducted a screening survey (n=1,000) targeting men and women aged 15-99 nationwide, and a main survey (n=210) targeting those who read web articles and have experience with generative AI, to understand the actual state of reading web articles after the spread of generative AI.
This survey revealed that the top reason for readers abandoning web articles was "unclear conclusion or answer" (44.8%), while "feels like an article written by AI/robot" accounted for only 17.1%. Furthermore, a cross-tabulation of respondents who answered "significantly increased/slightly increased" their AI usage frequency (n=137) and those who answered "unchanged" (n=53) confirmed that the group with increased AI usage had higher abandonment rates for almost all reasons for abandonment.
## Key Points of This Survey
- The top reasons for abandoning web articles were not AI-likeness (17.1%) but "unclear conclusion" (44.8%), "content duplication" (43.8%), and "strong advertising tone" (38.6%).
- It was confirmed that readers who use AI more frequently tend to be more critical of web articles regarding "information freshness," "content relevance," and "readability" (with a maximum difference of 19.9 percentage points).
- However, for "AI-likeness" alone, the difference between the increased AI usage group and the unchanged group was small at 3.2 percentage points, indicating that sensitivity to "whether it is AI-generated" itself did not change significantly even with increased AI usage.
## Information on the Detailed Report
The results of all 12 questions × all options not listed in this release, cross-sectional analysis of questions, and details of the questionnaire are available on the detailed report page below.
"Detailed Report on Web Content Reading Habits in the AI Era"
https://thinkmove.jp/survey/ai-contents-survey/
## Survey Overview
### Fact 1: The top reason for reader abandonment is not "AI-likeness"
When respondents to this survey (n=210) were asked (multiple responses) why they "abandon (stop reading)" web articles, the highest response was "unclear or vague conclusion/answer" (44.8%). This was followed by "same content repeated multiple times" (43.8%) and "feels like it's for advertising/affiliate purposes" (38.6%).
On the other hand, "AI-likeness," which has been discussed in the content industry in recent years, meaning "feels like an article written by AI/robot," was 17.1%, ranking 7th out of 9 options.
Source: Thinkmove Inc. "Survey on Web Content Reading Habits in the AI Era" (n=210, multiple responses)
All top three abandonment reasons are related to article structure, information design, and advertising expression, not the writing style itself. It can be concluded that readers abandon articles when they find it difficult to reach valuable information, such as "cannot reach the conclusion," "same content is repeated," or "the boundary between advertising and main text is ambiguous."
However, "being AI-generated" does not entirely have no impact on reader evaluation. In another question (Q8) of this survey, responses to knowing an article was AI-generated included "reading priority decreases, often search for another article" (12.4%) and "close the page" (4.3%), indicating that a certain number of readers are wary or abandon content specifically because it is AI-generated.
### Fact 2: Readers who use AI more frequently are more critical of web articles
This survey conducted a cross-tabulation of question Q11, which asked respondents "how their frequency of using AI for information gathering has changed compared to 1-2 years ago," and Q7 (reasons for abandonment).
Comparing the AI usage increase group (n=137) who answered "significantly increased/slightly increased" and the "unchanged" group (n=53), it was confirmed that the group with increased AI usage had higher abandonment rates for almost all reasons for abandonment.
Source: Thinkmove Inc. "Survey on Web Content Reading Habits in the AI Era" (cross-tabulation, multiple responses)
The largest differences were observed in "information is old/not updated" (+19.9 percentage points), "content was different from what I wanted to know" (+12.7 percentage points), and "difficult to read/hard to see" (+9.8 percentage points). Dissatisfaction regarding information freshness, relevance, and readability was significantly higher among the group with increased AI usage.
This result suggests that readers accustomed to instantly getting answers from generative AI may raise their bar for web articles, questioning "is there valuable information worth reading here?"
### Fact 3: Sensitivity to "AI-likeness" does not significantly change with AI usage frequency