The Power of Artificial Intelligence (AI) That Predicted 'Woo Young-woo Daebak' in Advance

KT is investing in successful content such as the drama "Extraordinary Attorney Woo" through its artificial intelligence box office prediction model. / Photo by Song Hyundo, Asia Economy intern reporter

[Asia Economy Intern Reporter Song Hyundo] It has recently come to light that the artificial intelligence (AI) developed by KT had predicted the craze for the drama "Extraordinary Attorney Woo" (Woo Young-woo) in advance. This was the result of applying an AI analysis model originally used in the financial business to the content industry. How was the AI able to foresee Woo Young-woo's success?

"Woo Young-woo" is a drama invested in by KT. It was produced by KT's subsidiary "KT Studio Genie" and aired through its own channel, ENA. Although Woo Young-woo started with a modest viewership rating of 0.9% for its first episode, it rapidly rose to become a so-called "national drama," recording a peak rating of 15.8% by the ninth episode.

The success behind this is attributed to the AI "content success prediction model" developed by KT. Released in March last year, this prediction model AI is now a core pillar of KT's content investment business. KT uses this AI to classify content they intend to invest in into 10 refined success grades and then pursues content production strategies according to their importance.

The key lies in the massive consumer data generated from KT’s Internet TV (IPTV) service. With about 9 million subscribers, the IPTV accumulates 700 billion customer request data annually. KT trains the AI with this data to build the success prediction model. When specific video content is input by synopsis, director, actors, etc., the AI classifies the expected number of viewers and revenue into 10 grades. The closer to grade 1, the higher the probability of success. The global streaming platform Netflix operates a similar prediction model.

◆ AI Predicts Content Success... Also Enhances Business Stability

KT Skylife, which operates the ENA channel, recorded sales of 254.2 billion KRW and an operating profit of 23.3 billion KRW in the second quarter (April to June) of this year. These figures represent increases of 45.2% and 0.4%, respectively, compared to the same period last year. Among these, advertising revenue reached a quarterly record high of 15.3 billion KRW, bringing the company one step closer to achieving its annual sales target of 1 trillion KRW. These results were made possible by Woo Young-woo’s high viewership ratings and the new viewers attracted thanks to the drama.

KT Skylife's sales in the second quarter of this year increased by 45.2% compared to the same period last year. / Photo by KT Skylife

The content business can generate significant profits when a "hit" occurs, but it is not a stable business with guaranteed returns. Due to the nature of content, where viewership ratings can fluctuate greatly depending on internal factors such as actors and scripts as well as external variables, content investment is naturally a cautious endeavor for investors.

"Woo Young-woo" was also a drama with many obstacles to success. The broadcasting channel ENA is a newly launched channel, and aside from lead actress Park Eun-bin, there were no prominent Hallyu stars in the cast. It was not a drama made by so-called "star producers" either. For this reason, director Yoo In-sik, who directed Woo Young-woo, expressed surprise in an interview with a media outlet last month, saying, "The drama is somewhat plain like Pyongyang cold noodles. I never imagined such enthusiasm from the beginning."

The content industry expects that success prediction using new technologies such as AI will invigorate the content investment industry, which is fraught with uncertainty. Netflix, in a research report published on its official blog, analyzed that "AI-based success prediction can improve the efficiency of marketing for dozens of in-house productions made annually" and "by minimizing 'human subjectivity,' which can negatively impact the success of some content, it enhances the stability of content management."