Only 40% of respondents believe the search results provided by ChatGPT are reliable. (Image courtesy of Nord VPN)
SEOUL, Feb. 28 (Korea Bizwire) – The advent of ChatGPT is anticipated to transform how we search for information online, yet a recent survey reveals that South Korean users still harbor doubts about the platform’s reliability. Specifically, only 40% of respondents believe the search results provided by ChatGPT are reliable.
According to the ’2024 Search Trends Report’ released by consumer data platform OpenSurvey, which conducted a mobile survey on February 14 and 15 targeting men and women aged 15 to 59 nationwide, the most commonly used platforms for internet searches are Naver (87%), YouTube (79.9%), Google (65.8%), Instagram (38.6%), Namuwiki/Wikipedia (34%), KakaoTalk (33.9%), Daum (29.1%), and ChatGPT (17.8%) in descending order, with Chat GPT’s ranking an indication that it has yet to secure a dominant spot in the search service market.
While 80.8% of internet users are aware of ChatGPT, only 34.5% have actually interacted with it, and a mere 18.3% continue to use it regularly. The report points out a conversion rate of just 42.7% from awareness to actual use, suggesting a significant drop-off.
The primary reasons cited for not using ChatGPT, or for discontinuing its use, include ‘no need or occasion to use it,’ accounting for 65.2% and 57.4% of responses, respectively. The report suggests that ChatGPT could improve conversion rates by providing clearer guidance on specific use cases.
Despite general satisfaction with the service’s responses, ChatGPT users express concerns over its reliability. Among those who have used the service, 66.7% find the search results or answers more satisfactory compared to traditional search services. However, only 40% regard these results as trustworthy.
Approximately 10.1% of ChatGPT users have opted for the paid version, ‘ChatGPT Plus,’ which costs $20 per month. Paid users rated their satisfaction and trust in ChatGPT higher than the average, scoring 4.26 out of 5 for satisfaction and 3.74 for reliability, surpassing the average scores of 3.76 and 3.33, respectively.
A team leader in the IT industry commented, “Compared to the free version based on GPT-3.5, the paid version powered by GPT-4 significantly outperforms in terms of convenience, productivity, and accuracy.”
Kevin Lee (kevinlee@koreabizwire.com)
