As technology rapidly advances, the application of Artificial Intelligence Generated Content (AIGC) in advertising is becoming increasingly widespread. AIGC uses AI technology to automatically create text, images, and videos, offering limitless possibilities for companies’ advertising efforts and setting a new trend in marketing.
Advertising, a highly creative and visually-driven industry, is gradually adopting AIGC technology to enhance efficiency and precision. This article explores the innovative applications of AIGC in advertising through examples of AI models like ChatGPT and DALLE 3 generating ad texts, images, and videos. Practice has shown that AIGC not only significantly improves the efficiency of ad production but also greatly enhances the accuracy and effectiveness of advertising. With these technologies, advertisers can more precisely reach their target audience, boosting the appeal and impact of their ads, bringing transformative change to the industry.
Artificial Intelligence Generated Content (AIGC)
AIGC refers to content generated using Generative AI (GAI) technology, rather than being created by human authors. It can automatically create large volumes of content in a short period. For instance, ChatGPT1 , developed by OpenAI, is a language model designed for building conversational AI systems that can effectively understand and respond meaningfully to human language input. Additionally, DALL-E 32 , another advanced GAI model by OpenAI, can create unique and high-quality images based on text descriptions in minutes, such as ‘mammoth on a desolate planet’ as shown in Figure 1. As AIGC technology progresses, more people believe it will become a new benchmark in the field of AI, having a profound impact on various sectors worldwide. From advertising to content creation, AIGC is revolutionizing the way we interact with information and media, ushering in a new era of efficiency, precision, and innovation.
Application of AIGC in Advertising Copywriting
Social media platforms like Facebook and Xiaohongshu have become an indispensable part of daily life. AIGC can automatically generate personalized ad copy based on users’ profiles and interests. For instance, on Xiaohongshu, ChatGPT can use human-provided ad templates to create high-quality, customized advertising content.
Application of AIGC in Advertising Images
DALL-E 3, developed by OpenAI, is an AI image generation model based on the Transformer 3architecture. Using DALL-E 3 model, advertisers can input product descriptions, target audience information, and ad style preferences to quickly generate high-quality advertising images. For example, as shown in Figure 3, if a restaurant owner wants to advertise a delicious steak, they just need to enter the food name and other details into the model. This allows them to generate a series of attractive restaurant food advertisements at a low cost. This approach not only saves production costs but also precisely targets the market, enhancing the appeal and effectiveness of the advertisements.
Application of AIGC in Advertising Videos
In today’s e-commerce landscape, content marketing is becoming increasingly diverse. Video content, with its vivid visual experience and rapid information dissemination, creates new opportunities for businesses. Consumers’ preference for video content drives the continuous growth of video creativity, making visual content an increasingly important bridge between consumers and products in advertising systems. However, compared to traditional text and image content, creating video content is much more challenging and costly. Producing a high-quality video requires professional skills, equipment, and time, leading to varying quality levels and difficulty in mass production.
With the continuous advancement of AI and Artificial Intelligence Generated Content (AIGC) technology, it is now possible to batch-produce high-quality video content using intelligent methods, bringing significant value to clients. AIGC technology not only improves production efficiency but also expands the boundaries of creativity, providing clients with more innovative and precise solutions. For instance, as shown in Figure 4, using video generation and editing tools based on diffusion models, the dynamic effect of rose petals falling near lipstick can be achieved, increasing customers’ purchase desire.”
Enhancing Efficiency and Precision with AIGC
The widespread use of AIGC has significantly accelerated the creative process in advertising. Traditionally, creating ad content required professional designers and planners to invest a lot of time and effort, especially when the ad needed to reflect the culture of the product, necessitating extensive research and design, which is time-consuming. However, with AIGC, advertisers can quickly obtain information related to product elements, swiftly distill elements that match the brand’s tone and target audience, and utilize them. Additionally, AIGC technology can generate various creative content including text ads, static images, dynamic images, and videos, greatly improving the efficiency and quality of ad production. Compared to traditional methods that require multiple revisions and refinements, AI-assisted production significantly enhances efficiency and speed.
In summary, ads generated by AIGC better meet the needs of the target audience, increasing the accuracy and effectiveness of ads. For example, by analyzing user data, AIGC can generate personalized ad content, boosting user engagement and conversion rates. Notably, large companies like Alibaba have begun exploring the integration of models like ChatGPT and DALL-E 3 to achieve fully automated e-commerce ad creation and placement, ushering in a new era of e-commerce advertising [1].
Future Prospects and Challenges
Recently, in an interview Lingyishuke CEO Jian Feng mentioned that design team is the first to benefit from AIGC software. After introducing AIGC, the workload that previously required at least five designers a week to complete can now be finished in two to three days. Furthermore, AIGC can create virtual hosts or presenters for advertising in the near future, providing immersive experiences and increasing user interaction. AI-automated content generation drastically reduces the technical time required for ad creation. AIGC can also quickly optimize and adjust ad content based on creator feedback. By using AIGC for ad creation, not only is time saved, but ad creators can focus on creative ideas and content exploration, producing more impactful and influential ads.
However, despite the immense potential and advantages of AIGC, its output heavily depends on human-provided data resources, which can lead to training defects and copyright infringement risks. AI may fabricate information when dealing with complex issues, leading to the misuse of cultural elements, damaging the essence of ads, and harming the brand image. Additionally, AI has a probability of generating seemingly reasonable but incorrect answers when handling complex, ambiguous, or open-ended questions. When ad creators use AI to handle related creative content, if they fail to detect errors in cultural element stitching, misinterpretation of cultural customs, or inappropriate handling of sensitive culture, it could lead to the misuse of traditional elements due to the limitations of AIGC. Such misuse could not only destroy the cultural essence of the ad but also cause irreparable damage to the brand image.
In conclusion, with the continuous development of AIGC technology, its application in advertising shows tremendous potential and advantages. Although there are data dependency and potential copyright issues, by exploring more reasonable usage methods and establishing appropriate regulatory mechanisms, AIGC is expected to further drive breakthroughs and transformations in ad production on the existing foundation, continually optimizing its application effects.
References
[1] 尹思筠.AIGC應用於“國潮”風格廣告創意中的優勢與困境[J].北京文化創意,2024,(02):60-65.
[2] 砍柴網（2024）。過去一年,他們如何靠AIGC搞爆款廣告?
The work described in this article was supported by InnoHK initiative, The Government of the HKSAR, and Laboratory for AI-Powered Financial Technologies (AIFT).
(AIFT strives but cannot guarantee the accuracy and reliability of the content, and will not be responsible for any loss or damage caused by any inaccuracy or omission.)
- ChatGPT is a large language model based on the GPT-4 architecture. It has been upgraded and optimized using vast amounts of data from the internet through techniques such as fine-tuning and reinforcement learning based on human feedback. ↩︎
- DALL-E 3 is a large image generation model that uses a diffusion model architecture and techniques like language-image contrastive learning to create images from text descriptions. ↩︎
- The Transformer model is based on the self-attention mechanism, allowing it to consider all elements simultaneously when processing data. Many large AI models today are built on the Transformer architecture. ↩︎
