With the rapid development of AI technology, AI advertising films are gradually becoming the new favorite of brand promotion. However, the production and application of AI commercials have not been smooth sailing, facing technical, legal and ethical challenges. This article delves into the current state of AI commercials, analyzing which brands and production teams are actively using AI technology, as well as the difficulties and solutions they encounter in practice.
To what extent has AI penetrated the content of advertising films? The answer may be beyond many people’s expectations.
According to the “2025 China Advertiser Marketing Trend Survey Report” released by CTR (CCTV Market Research), 53.1% of advertisers will use AIGC technology in creative content generation, which is higher than data analysis, customer service and other links, of which nearly 20% of advertisers need to use AI in video creation.
But can using AI really help advertisers reduce costs and increase efficiency? The situation does not seem so optimistic: in order to ensure that the protagonist in the AI advertisement does not “change his face in the middle” of the film, repeated generation and debugging dozens or hundreds of times is commonplace; In terms of process, Party A’s demand and AI output are still separated by 100,000 thousand miles; Even if the AI film is finally generated that is indistinguishable from true and false, users may not necessarily pay for it – before the Toys”R”Us AI advertisement was criticized as “soulless” and “insulting IQ”, and then there were KFC and McDonald’s old enemies who both overturned, and some brands fell into legal and ethical disputes. In April this year, director David Blagojević’s viral advertising short film using AI for KFC was launched and fell into huge controversy. Many directors and photographers have accused David Blagojević of “plagiarizing” his footage and ideas in the name of AI, and photographer Alejandro H. Madrid has criticized the former for “no creativity”.
It is undeniable that AI has a positive effect on improving efficiency and reducing costs. However, it is difficult for all parties to agree on the use of AI, and the effect of the film is also erratic. There are a series of technical hurdles and legal and ethical disputes in the front, and then there are various stakeholders wrestling with each other, where will the AI advertising film go?
Who is filming? Big brands are enthusiastic, and AI studios are taking advantage of the momentum
AI advertising films have been quite common in recent years, and enterprises and brands in the fields of catering and food, FMCG, automobiles, digital 3C, home appliances, gold jewelry, and the Internet are trying this technical means.Overall, big brands are made more frequently,Such as McDonald’s, Lenovo, Coca-Cola, etc. are veteran players in this area.
For example, Lenovo, which launched the brand short film “Born in 1984” to celebrate its 40th anniversary last year, as well as the advertising film “Our Workstation” to promote the new products of the Zhaoyang series earlier, and the advertising film “Unbounded Elegance” to promote the YOGA Pro 14 Ryzen AI Yuanqi Edition, etc., all use AIGC technology to varying degrees.
B-end product manager’s ability model and learning improvement
The first challenge faced by B-end product managers is how to correctly analyze and diagnose business problems. This is also the most difficult part, product design knowledge is basically not helpful for this part of the work, if you want to do a good job in business analysis and diagnosis, you must have a solid …
View details >
The former joined hands with a group of leading anchors to bring a series of AI short films in 618 last year, and the latter has released AI advertisements such as “Domestic Products Have Their Own Sources” and “Second-hand is Not Expensive” in the past two years in 618, Double 11, 815 and other promotional activities.
Lenovo’s 40th anniversary brand short film “Born in 1984” produced by AIGC
The reason why big brands are more willing to produce AI short videos is closely related to their own resource advantages and communication demands.
On the one hand, leading brands and large manufacturers have a large amount of advertising and sufficient budgets, and can withstand a certain degree of innovation risk and allocate some resources to try AI advertising films. Big brands need gimmicks, they need to reflect their own pioneering temperament, AI is in the limelight for a while, and brands certainly hope to use their popularity to add points to their image. It is not difficult to understand why the Internet and technology giants are particularly enthusiastic about AI short films.
On the other hand, it is becoming more and more difficult for traditional TVC films to shoot new ideas, and the competition of leading brands is fierce, and AI needs to provide new stimuli and open up creative ideas. For example, the low operational threshold of using AI to generate videos allows brands to play with some new ideas in interaction, and CATL once launched an AIGC interactive advertisement during the Guangzhou Auto Show, allowing users to enter basic information and upload photos, and AI can generate an exclusive personalized customized short video to stimulate users’ enthusiasm for interaction.
CATL has invited users to shoot AI advertisements together to open up new ideas for interaction
The enthusiastic pursuit of brands has undoubtedly boosted the outlet of AI advertising films, bringing new opportunities to advertising companies and downstream agency producers and AI studios.
Brands will not use AI to generate advertisements themselves, and they still have to rely on downstream partners for specific creativity, production, and distribution. Almost without exception, leading advertising companies are actively embracing AI, including cultivating professional AI creation teams, independently developing or tuning AI tools.
WPP Group hugged the thighs of Silicon Valley giants, cooperated with Microsoft to use tools such as Azure AI Vision to improve video production effects, and joined hands with Google to call the Gemini model and the self-developed “WPP Open” system integration to optimize AI video narration generation and other functions; Publicis uses AI combined with the optimized technology of eye tracker technology to produce more than a dozen video materials adapted to different platforms a day, and customers such as GlaxoSmithKline have begun to use these services; Huayang Lianzhong launched the video processing tool “Lingjiang” last year, which can intelligently generate mixed and edited video content.
The momentum of the start-up team is not as huge as that of the head advertising company, but it has also won the favor of some leading brands by taking the route of intensive cultivation. For example, the Zhizhu Ai team that helped FAW Toyota produce the “first automotive commercial AI video in China”, the creative team of Xixifus who produced AI short films for the Kuaishou Film Festival, the breathe studio that cooperated with Anta, and the heterogeneous culture studio that cooperated with brands such as Suntory and Dong’e Ejiao are all rising stars worthy of attention in the field of AI advertising films in recent years.
In contrast, there are not many cases of AI advertising films related to small and medium-sized brands on the market, which is inseparable from limiting factors such as cost and technology.Different from large brands with strong funds, small and medium-sized brands do not have much budget, and their own brand appeal and popularity are limited, even if they catch the hot spot of AI, it is difficult to get the heat bonus to match the head brand, and they are naturally more inclined to traditional production forms with guaranteed quality.
Perhaps only when the production threshold of AI advertising films is further lowered and the effect has withstood the test of time will more small and medium-sized brands join in – however, judging from the current difficulties and development trends of AI advertising films, there is still a long way to go before this day.
Some people overturn and some go out of the circle, and there is no universal formula for AI commercials
At present, there are two main types of AI advertising films on the market: one is the full-process AI short film from creativity, storyboard script to final film, and the other is the use of AI tools/technology in some links, mainly used to deal with certain shots and scenes that are difficult to shoot.
The former solution is technically difficult and uncertain, and big brands only use it as a means to try it out, usually used to compete for a “industry first right” to show technical advantages, or pay tribute to the classics to express feelings. For example, Taobao launched the “industry’s first all-AI-produced” Taobao Creation Day promotional video as early as 2023, and Coca-Cola released a Christmas video that replicated the classic advertisement “Holiday is Coming” during Christmas last year.
The latter solution is more flexible and more popular, and many first-tier brands and Internet giants have launched corresponding works. Suntory’s Spring Festival short film and Anta’s “Irving Generation” new product promotional video take the form of a combination of AI and real shooting, with real actors appearing on the scene and using AI to assist in synthesizing some scenes, which can not only save the cost of external shooting, but also reduce the impact of objective factors such as venue rental schedule and weather.
Taobao has tried the whole process of AI production as early as 2023
However, it can be seen from existing cases that no matter which solution is adopted, it is impossible for AI commercials to completely leave the control of people, and it is unrealistic to generate AI autonomously in the whole process.
In terms of production process, different AI tools have different characteristics and advantages, and they need to rely on the experience of creators to choose the right tool. Although the 2024 AI promotional video of the Grevia 4WD version produced by the newly established AI production team Zhizhu Ai for Toyota is entirely generated with AI, the choice of tools still depends on the team members: Midjourney for background and character generation, and Runaway, Deforum, Animatediff, etc. are used for some clips. Which tools are more suitable for which link depends on the team to explore the rules in practice.
The shortcomings of AI, such as unstable picture generation and easy face change, are inseparable from human post-adjustment. In the short film cooperated by the Hee Xifus team and Kuaishou, the character “changed face” problem, and the appearance of the protagonist in the film could not be consistent, and after countless attempts, he found the only version that met the standard – the team couldn’t help but ridicule, “Just like Doctor Strange said that out of 14 million futures, only one will win.”
In addition, AI understanding capabilities are also flawed, and it may not be able to read the style that the brand wants, or properly present brand elements, and many details need to be artificially refined.
Xunguang Culture and Heterogeneous Culture once reviewed the process of making AI short videos for Dong’e Ejiao on Shuying.com, and mentioned this issue. For example, when directly telling the AI that it wants historical and cultural elements, youthful and relaxed tones, and asking it to “naturally reveal” the brand logo, the AI often fails to accurately recognize it. In particular, AI’s understanding of “realistic style” stays on the surface, and it is easy to generate unreasonable pictures by automatically substituting scenes and characters into realistic references. For example, brand elements such as “Ejiao Well” and “Ejiao Ancient Town” in the film should be three-dimensional modeled first, and then processed with AI, and three-dimensional construction and real-life synthesis should also be used when making transition shots.
Dong’e Ejiao advertising film made using AI
The bigger problem is that even if there is a creative team behind the scenes, it cannot solve all the technical defects of AI advertising films at this stage.Some controversial rollover cases in recent years are enough to illustrate the problem.
The most obvious technical problems occur in the generation of character dynamics and action scenes.Due to the limitations of AI’s ability to understand complex scenes, the more storyboards and actions involved in the commercial, the more likely the generated scenes are to appear incoherent. When it comes to complex dynamics such as multi-character interaction and rapid switching of scene actions, AI’s detailed processing is often unsatisfactory due to reasoning capabilities.
A typical case is that McDonald’s Japan launched an AI advertisement last year to promote the French fries promotion, in addition to the basic error of the AI portrait having one more finger, the entire commercial was also criticized for the character’s unnatural expression, lack of affinity, and inconsistent actions, and even some netizens ridiculed “looking like an advertisement shot by KFC undercover”.
McDonald’s AI commercial film in Japan, character expressions and actions were criticized as “weird”
In addition to technical injuries, the deviation of the promotion direction of the brand and production team and the ensuing pressure of public opinion have also made the effect of AI commercials more uncontrollable.There are two common public opinion controversies: one is to over-exaggerate technical advantages and advocate the power of AI; second, it only pursues technology and pays attention to pictures, ignoring the communication of content creativity and brand sincerity.
You know, one of the reasons why humans love and fear AI is that they are worried about being replaced by AI. If brands blindly promote the power of AI and emphasize the advanced nature of AI, it will invisibly amplify this fear and cause public opinion to resist.Don’t forget that the fundamental value of TVC is not to promote technology, but to convey brand concepts and shape brand tone, and its ultimate purpose is to release brand value to users and establish emotional connections. If the commercial is reduced to a tool to promote AI, the empty technology lacks creativity and sincerity, which is tantamount to putting the cart before the horse.
The most controversial recent negative case is the AI short film launched by Coca-Cola on Christmas Eve last year. Coca-Cola revealed that the 30-second advertisement was completed by three top AI studios in the industry, using four artificial intelligence models to highlight technical advantages. Objectively speaking, the quality of the commercial itself is not too bad, and the restoration of Christmas elements such as hot air balloons, polar bears and elk running scenes is still qualified. However, after the launch of the short film, the biggest dissatisfaction of netizens is that the whole short film lacks theme and substantive creativity, and the cold AI-generated scenes and characters are not reminiscent of the warm Christmas atmosphere.
Coca-Cola’s Christmas AI short video was criticized for lacking temperature
It is true that public opinion has always been elusive, and the phenomenon of traditional advertising films overturning is also common, and it is even more difficult to find the “high score formula” for AI advertising films that have not been around for a long time. But fortunately, by summarizing the experience and lessons of positive and negative cases, practitioners have gradually explored some rules.
First of all, at any time, the most important thing in brand TVC is “intention”. A good commercial – whether it is traditional manual shooting or AI-assisted production, the first thing is to convey “human” care.
Some brands and production teams that truly understand consumers and the meaning of commercials have long learned to use the sense of conflict and contrast brought by AI to show the “human” side of the brand. Like Dove, while promising never to use AI to generate characters in advertisements, while teaching users how to use AI correctly, get rid of AI anxiety, and win the hearts of the people in the short film “The Code” and a series of reports; Another example is Meituan’s “Reunion 2024” New Year’s short film, which actively embraces AI in production, but the plot theme is to tell people that although AI can shoot the Spring Festival Gala in 3 minutes and instantly make a Chinese New Year’s Eve dinner, it will never replace the sincere communication between relatives.
Meituan’s AI short film conveys human care through the plot
The second is to leverage the strengths and avoid the weaknesses in technology,Knowing that AI has flaws in the processing of dynamic scenes and character expressions, the creative team has learned to avoid complex actions, reduce close-ups of characters’ faces, and focus on the advantages of AI in scene generation, and work picture quality and theme creativity.
Coca-Cola actually had a successful experience before it overturned. Its earlier “Masterpiece” short film conveys the creative plot of Coca-Cola through famous paintings in the museum, shows the process of bottle transformation, and combines Coca-Cola with “Shipwreck”, “Scream”, “Girl with Pearl Earrings” and other artistic masterpieces, conveying the brand concept of “Real Magic”, which has received excellent reviews.
Coca-Cola’s earlier “Masterpiece” AI commercial was highly praised
In addition, short videos that fit the brand’s own tone, and the down-to-earth performance style is easier to eliminate the sense of distance between people and AI.
For example, abstract style AI short films can reasonably convert characters’ unnatural expressions, slow movements and other hard injuries into “laughs”.
Ctrip’s AI advertising film “Egyptian Pharaoh Lying Flat All Over the World” last year’s Double 11 is a good example, shaping the majestic and mysterious Egyptian pharaoh into a migrant worker who “escaped” from the pyramid and “lying flat” in major attractions around the world.
Ctrip’s Double 11 AI short film “Egyptian Pharaoh Lying Flat All Over the World” is out of the circle because of its abstract painting style
It can be seen from the above cases that it is indeed not easy to fully understand the preferences of public opinion and overcome technical problems. However, as long as the brand and creative team are attentive enough, they can still shoot high-quality AI advertisements that satisfy users under limited conditions.
The road ahead: one big opportunity, two big challenges
What is certain is that AI commercials are on the road of becoming more and more mature.
Next, as various tools continue to innovate and production experience continues to accumulate, a series of technical barriers are expected to be broken through one by one.
In terms of technical infrastructure, AI products/tools on major platforms are constantly improving: Kuaishou Keling recently released the 2.0 video generation model, which has made many breakthroughs in dynamic quality, semantic response, picture aesthetics and other links, such as the new Multi-modal Visual Language interaction mode, which can improve AI’s understanding of creative instructions and further improve processing efficiency; Byte’s Dream recently opened the internal beta of the video model version 3.0, which not only provides richer lens language and scene styles, but also optimizes the character and picture generation effects. For example, in terms of picture effects, it not only supports the generation of 8K resolution ultra-clear images, but also skillfully uses high-end imaging techniques such as depth of field lenses and light and shadow levels.
The technical routes chosen by AI products such as Keling and Jimeng have their own focuses, but one thing is the same: make efforts to differentiate and continue to optimize and iterate their own advantageous functions. As mentioned earlier, considering the limitations of R&D investment and technology, it is difficult to find an all-round AI tool on the market at this stage. However, as long as the big studios find their own advantages and continue to amplify and deepen them, and the production team chooses the right tools based on their own experience, they can also significantly improve the quality of the film. For example, the ability to interact with and recognize commands is of great benefit to improving production efficiency; The optimization of the dream lens generation effect can provide the production team with more stylized options, so that AI advertising films can move in the direction of “thousands of pieces”.
Major manufacturers are working hard to optimize AI products
Of course, there are still obstacles on the way to the development of AI advertising films – especially legal and ethical disputes, which are also the two biggest difficulties for brands and advertising companies.
On the one hand, brands are worried that feeding brand materials to AI will expose brand equity and be used by other teams and competitors; on the other hand, they are more worried about AI content infringement, falling into legal disputes, and affecting brand image. This puts even more pressure on advertising companies caught between Party A, users, and regulators. A report by the World Federation of Advertisers (WFA) in the second half of last year pointed out that 80% of advertisers are “seriously concerned” about the ability of advertising agencies to use AI technology, with the main concerns being legal, ethical and goodwill risks.
In the short video of the 70th anniversary of Volkswagen Brazil, the use of AI technology to “resurrect” the late singer Elis Regina caused a lot of controversy: some netizens praised the full score of feelings, and the way Elis Regina and her daughter reunited in singing was very touching; Some netizens and media have also criticized AI’s resurrection of deceased well-known figures as potentially misleading to the public, and Brazil’s National Advertising Self-Discipline Commission (Conar) has launched an ethics investigation into the public.
Volkswagen’s use of AI to resurrect the late singer Elis Regina in the commercial has drawn criticism
Objectively speaking, due to the fact that the rise of AI advertising films has not been long, and the national conditions of various countries are different and the interests of all parties are complex, it is difficult to form a unified norm in the industry at this stage. Taking the copyright ownership and infringement liability determination of AI advertising films as an example, the European Union, the United States and other places have completely different regulations.
Fortunately, after a period of exploration, brands and advertising companies have accumulated a lot of experience, even if they cannot completely resolve legal and ethical disputes, at least they have gradually found the right boundaries and avoided risks to the greatest extent.
For example, for copyright and privacy protection issues, large advertising companies have solved them through privatization deployment and self-developed tools. Like WPP’s “WPP Open” system, although it integrates large models from Amazon and Google, it maintains the security of its own intellectual property library. It is worth mentioning that for advertising companies, this ability to control AI technology has another important significance: making customers realize that AI cannot completely replace people, proving that advertising agencies still have irreplaceable value in the AI era.
In the face of more complex intellectual property ownership and ethical disputes, the current focus of all parties is to seek common ground while reserving differences, hoping to reach a consensus in the industry and strengthen self-restraint. For example, iFLYTEK launched the “virtual human identity authentication system” for digital human copyright protection earlier, which was well received, and in time, similar copyright protection systems for materials and advertisements are expected to be promoted on a larger scale.
In the final analysis, no technological change can be achieved overnight, and detours and controversies have always been an indispensable part. But only by exposing and solving problems in practice can the industry clear the fog and find the right direction forward.
