From a woman's childhood to old age, memories shared with Gyeonggi Province unfold in a sentimental video lasting 30 seconds. Lovers whispering sweet nothings beneath the night sky become parents of a child, enjoying the warm sunlight in a park. This video is a promotional film created in just seven days using generative artificial intelligence (AI) by internal workers of the Gyeonggi Tourism Organization, without a professional production company or actors. Compared to traditional methods that required a budget of tens of millions of won and production time of several months, this dramatically reduces expense and time.
In this way, artificial intelligence (AI) is shaking up the very production methods of the advertising industry. In the past, it cost tens of millions to hundreds of millions of won to produce video advertisements, but now, all that's needed is AI to complete high-quality video content in a matter of hours.
◇ 86% of U.S. advertisers plan to adopt AI in advertising…90% production expected next year
According to the '2025 Digital Video Advertising Strategy Report' by the Interactive Advertising Bureau (IAB) on the 5th, 86% of advertisers in the U.S. said they are currently using or plan to adopt generative AI. By 2026, it is expected that 90% of all advertisers will utilize AI. The report predicted that about 40% of all digital advertisements will be produced by actual AI.
The backdrop for these changes is that global big tech corporations like Google and Adobe are rapidly accelerating the commercialization of AI-based advertising technology. These corporations, which have been developing video generation technology using AI for several years, have recently optimized functionalities for advertising practices and intensified platform integration.
On the 31st of last month (local time), Google introduced a lightweight version of its generative video model 'Veo 3,' specialized for the advertisement production environment, called 'Veo 3 Fast.' This model automatically generates 1080p high-definition video and audio in minutes with just one image and text, allowing for ad content production at a low expense of about $0.40 per second. Google has integrated this across its platforms including YouTube, Display Network, and search advertisements, allowing advertisers to handle everything from video production to distribution in a one-stop process through API.
Adobe also added AI-based video advertisement generation capabilities to its marketing automation platform 'GenStudio' on the 29th of last month (local time). This feature automatically produces dozens of ad videos with just product images and key phrases and can be deployed in real-time to major platforms like Google, Amazon, and Meta. It supports not only simple video generation but also the production of various versions of content and A/B (advertising performance comparison) testing.
In fact, changes utilizing these technologies are rapidly occurring in the advertising industry. An advertisement for the online trading platform 'Kalshi,' which appeared during NBA Finals streaming on June 11, is a leading example created by AI in just two days. This 30-second video, composed of bizarre yet eye-catching scenes like a bride fleeing from a wedding, a farmer lounging in a pool of eggs, and an elderly person playing with an alligator's head, was completed for just $2,000 (about 2.76 million won). This is only 5% of the cost of traditional advertising production. The advertising producer PJ Accetturo noted on his X (formerly Twitter) that they selected 15 scenes after 300 to 400 attempts with AI.
In Korea, the barriers to entry for advertisement production are also being dismantled. LS Group produced a promotional video for its power system using AI, achieving both video and sound without filming while significantly reducing production expense and time. LG H&H created an advertisement for laundry detergent by recreating table tennis player Shin Yu-bin with AI, implementing various scenes through AI image synthesis instead of shooting with a real model. Seoul Milk showcased an advertisement that restored actress Park Eun-bin's childhood using AI, cross-cutting between her childhood and current appearances to enhance consumer attention.
◇ AI instead of cinematographers and actors…consumer backlash is a challenge
The industry analyzes that it is now time to redefine the role of humans across the entire advertising ecosystem. While the roles of planners, cinematographers, actors, lighting technicians, and other existing staff are diminishing, the role of controlling AI and designing strategies will inevitably become prominent.
In the case of Google Veo, it has learned from vast content such as YouTube and Google Images, enabling it to automatically orchestrate camera movements, lighting, and character motions with just text, and is integrating this into its advertising platform so that advertisers can directly request video production via API.
However, this trend is not welcomed by everyone. Consumers are expressing a peculiar sense of discomfort and unease regarding advertisements produced by AI. According to a survey by market research firm Nielsen in January of this year, 69% of global consumers said they can recognize AI-generated content, and among them, 55% responded that it is "emotionally unpleasant."
A representative backlash example is the AI model advertisement for the clothing brand Guess in the August 2025 issue of the U.S. edition of Vogue. Criticism arose over the advertisement featuring a blonde AI model "VIVIEN," questioning, "Should we be compared to non-existent people?" This comment received over 60,000 'likes' on TikTok. Advertising producer Seraphinne Vallora clarified that "the AI model was generated based on the poses and body types of real models, and real models are also used concurrently."
Lee Eun-hee, a professor of consumer studies at Inha University, noted that "since AI models are not real individuals, consumers lose the basis to judge the authenticity of the advertising message," and added, "particularly, the manner in which virtual models that have no entity mimic humans can trigger feelings of being deceived or misled in consumers, potentially intensifying the 'uncanny valley' phenomenon." He further stated, "Currently, there are no clear regulations in advertising law regarding the use of AI models, so institutional improvements are urgently needed from a consumer protection standpoint."