생성형 AI 서비스로 만들어진 불확실한 정보, 저품질 콘텐츠들이 기하급수적으로 증가하고 있다. 이에 사용자들의 편의성을 해치는 검색 서비스의 질적 저하가 심화되고 있다. e4ds news 취재 결과 네이버, 구글 등 주요 검색 포털에서 정보를 검색할 때 생성 AI 기반 블로그 콘텐츠들이 난립하고 있는 것이 확인됐으며, 검색 포털 사용자들의 불편도 증가하고 있는 것으로 나타났다.
.jpg)
▲Part of a post titled 'These days, the serious portal search situation'. Search users are experiencing increasing inconvenience due to low-quality content generated by AI. / (Capture-theqoo)
AI posts low-quality, obvious information and fake information
AI Blog Monetization and Automation, the Main Cause of Ecosystem Destruction
The amount of uncertain information and low-quality content created by generative AI services is increasing exponentially. As a result, the quality of search services is deteriorating, which is damaging user convenience.
According to e4ds news coverage, AI-based blog content is proliferating when searching for information on major search portals such as Naver and Google, and the inconvenience experienced by search portal users is also increasing.
A post titled 'These days, the portal search situation is serious' is gaining sympathy from netizens in online communities. The author searched for Itaewon World Food Recommendations on Naver and checked the ‘restaurant name, address, signature menu, price, etc.’ in a blog post. However, the restaurant in question was not only not found on Naver Maps, but the restaurant name that came up through the address was completely different from the one in the blog post.
The author said, "It turned out that it was a list of restaurants that had been made to look good by using generative AI like Chat GPT," and pointed out, "What's even more creepy is that the other articles were also filled with recommendations for non-existent restaurants that were also generated by AI."
In addition, he expressed concern that the false AI information on search portals is not limited to restaurant information, but is a serious problem with blogs that list 'plausible but useless information' while occupying the top positions in all areas of life knowledge.
Netizens responded with comments agreeing with the spread of false information and low-quality information through AI content, such as, "These days, YouTube has a lot of these AI misinformation videos", "These days, blogs and intellectual sites are all filled with AI posts, making it really hard to search", "○Story is 99% AI posts", "Cafe's are even worse; even the posts and comments are all empty conversations, as if they were made by AI", and "Regulations are needed, maybe hiring monitoring agents".
■ AI false posts are rampant on search portals, such as ‘XX comparison’ and ‘XX area accommodations and restaurants’ 
<br /> ▲There are an increasing number of articles in search results that are suspected to have been generated by AI.
When you search for 'Hanok Stay in Hwagok-dong, Seoul' in the Naver search bar, the article 'Hanok Stay in Hwagok-dong, Gangseo-gu, Seoul: Recommendations for emotional accommodations with good cost-effectiveness' is at the top of the search results. Below this is 'Recommendations for emotional hanok stays in Magok-dong, Gangseo-gu, Seoul and price information', and all of these articles are ultimately determined to be AI posts containing false information.
First, when you enter the first article, you will see ‘The charm of Hwagok-dong Hanok Stay’ which is not related to the post title, followed by a price comparison table showing the best value accommodations such as ‘Hanok guesthouse 50,000 won per night’ and ‘Hwagok Hanok accommodation 70,000 won per night’.
Even when searching for the business name, which did not seem to be a specific business name, on Naver Map and the registered business portal, I could not find the accommodation facility. The article about Magok-dong Gamsung Hanok Stay was the same. These articles were excessively filled with banner ads on the left and right sides and banner ads inserted between paragraphs.
The search terms that the reporter found that are frequently suspected of being generative AI posts are ‘Tteokbokki calorie comparison’, ‘XX region travel accommodation comparison’, ‘Domestic private hot spring hotel’, and ‘Causes of shoulder pain symptoms’. It is felt that the frequency of observation is high when they are combinations of specific keywords.
■ AI Suspicious Posts, Obvious Words, False Information, Repeated Similar Structures... The Criminal of 'AI Blog Monetization' .jpg)
▲The frequently asked questions structure is filled with repeated, obvious questions and answers that are somewhat less helpful.

▲In the related comments, comments suspected to be from comment-generating bots continue to appear.
When analyzing suspected AI posts for our paper, the commonalities observed were: low context between title and content, obvious answers with low specificity, false information, and repetition of similar structures.
Local accommodation recommendations provide general information about the area, tips for choosing accommodation, types of accommodation, etc., and sometimes include a list of accommodations. However, the accommodation listings are often false, and it is easy to find cases where the business name, price, and address are false. In addition, it was confirmed that the comments on the posts were patterned and had similar writing styles.
Suspicious AI posts often include structures related to △comparison △recommendation △ranking, and many of them add Q&A structures such as ‘Top 5 frequently asked questions about XX’ at the bottom of the post.

▲Numerous blog contents introducing AI blog monetization and automation
It is pointed out that the proliferation of low-quality content is being done by those who aim to generate advertising revenue through views. If you search for AI blog monetization and automation in the search bar, numerous related methodologies are introduced, and related seminars and lectures are also promoted.
If the search ecosystem loses user trust, this can also have a big impact on the digital and AI competitiveness of the portal. As false information and meaningless low-quality content become widespread, fatigue and departure from portal use are bound to increase. Accordingly, users’ voices are also growing louder for portal content regulation and monitoring.