In an exciting announcement on May 28, 2023, NVIDIA and WPP joined forces to forge a groundbreaking alliance, unveiling their ambitious plans to build a cutting-edge content engine. This innovative collaboration harmoniously merges the immense potential of NVIDIA Omniverse™ with the prowess of artificial intelligence technology. The result? This will create great opportunities for advertisers to produce content faster, cheaper and at scale. The advanced content engine aims to enable them to deliver more personalised and tailored content to the right audiences, helping them to achieve their marketing goals and growth.

Get ready to delve into AI-generated advertising, where algorithms craft captivating ads tailored to individual viewers. In this article, we’ll explore the potential impact of AI-generated ads on video advertising, discussing advantages, limitations, and real-world applications that will shape the future of advertising.

AI-Generated Advertising: Redefining Content Creation

AI-generated advertising transforms the brand-audience relationship, employing advanced artificial intelligence technologies such as machine learning and deep learning algorithms. Unlike traditional methods, AI-generated advertising employs data analysis and pattern recognition to create customised content tailored to each viewer. By dynamically adapting and optimising content in real time, AI-generated adshave the potential to deliver personalised, relevant experiences at massive scale based on user behaviour and contextual data. This targeted approach drives enhanced engagement and conversion rates for advertisers.

According to a report by MarketsandMarkets, the market for generative AI is expected to grow significantly, with an anticipated increase from USD 11.3 billion in 2023 to USD 51.8 billion by 2028. This demonstrates the growing recognition and investment in the potential of AI-generated advertising.

The technology behind AI-generated ads

AI-generated advertising is driven by a wide range of technologies and algorithms. Let’s take a look at some of the most notable of them:

Generative Adversarial Networks (GANs)

GANs have revolutionised generative AI by introducing a competitive framework between the generator and discriminator networks. They have demonstrated remarkable capabilities for generating realistic and diverse content across various domains. Brock, Donahue, Simonyan (2018)  showcased the generation of synthetic photographs with their technique BigGAN in their research paper titled  “Large Scale GAN Training for High Fidelity Natural Image Synthesis”. The result is an amazing set images that are practically indistinguishable from real photographs.

Variational Autoencoders (VAEs)

VAEs have proven to be powerful generative models, capable of learning and generating new instances that capture the underlying distribution of the training data. They have been extensively used in applications like image generation and data augmentation. As can you can see below, GitHub user wojciechmo had some great success in generating human-looking faces with VAEs.

Transformers

While they are less used in video content creation, Transformers have greatly influenced natural language processing and generation tasks. Transformers’ attention-based architecture enables them to capture long-range dependencies effectively, making them invaluable in tasks such as language translation, text generation, and chatbot systems.

Autoregressive models

They have shown impressive results in generating high-resolution images and audio samples by modelling the conditional probability of each element in a sequence based on previous elements. Autoregressive models have pushed the boundaries of generative AI in terms of visual and auditory fidelity. Take a look at the example below from Reza Fazeli of some celebrity faces generated using autoregressive models.

Reinforcement Learning

Reinforcement Learning (RL) in Generative AI: RL algorithms, combined with deep learning techniques, have been applied to generative AI, enabling agents to generate novel and creative content. RL has contributed to areas such as artwork generation, music composition, and game-level design.

These are just a few examples of the prominent generative AI technologies that have gained recognition and made significant contributions to the field. The field of generative AI continues to advance rapidly, with new techniques and models being developed regularly such as Google’s recent release of their state of the art PaLM 2 language model.

The Rise of AI-Generated Advertising

According to the 2022 State of Marketing and Sales AI Report, 77% of marketers have already incorporated AI automation to a certain extent, albeit in less than a quarter of their tasks. Looking ahead, approximately 74% of marketers anticipate that over a quarter of their tasks will be intelligently automated within the next five years. This projected growth highlights the increasing reliance on AI-powered automation and generative AI in marketing operations. As a result, AI is expected to become an integral and widespread component of marketing stacks shortly.

At the moment, while AI is a worthy platform with many possibilities, clear control and guidance from marketing specialists are indispensable. However, this may change soon enough. Let’s take a look at a few sensational AI-generated ads that have shaken up the Internet.

“Pepperoni Hug Spot” – AI generated pizza commercial

This is a remarkable example of an AI-generated ad that caused a stir, and even captured the interest of Elon Musk, replying to a Pizza Hut comment about the ad on Twitter:

The advertisement for pizza was brought to life through a remarkable blend of AI technologies. Utilising Runway’s Gen-2 for video footage, GPT-4 for the script, Midjourney for still images, Eleven Labs for voice-over, and Adobe After Effects for the final assemblage. The entire commercial was astonishingly crafted in a mere three hours using AI software.

The pizza commercial features a fictitious pizza joint named “Pepperoni Hug Spot.” It showcases individuals enjoying their mouthwatering pizzas, with AI-generated visuals adding a touch of excitement and creativity.

Examining this AI-generated pizza advertisement reveals the immense potential that AI algorithms hold. While the current visuals are somewhat “disturbing” at the moment, given that this is only one of the first attempts at AI-generated ads, the overall outcome is captivating and shows with a bit of fine-tuning we could see genuine video ads being generated by AI in the not-too-distant future.

But, the big question for advertisers – is the Pepperoni Hug Spot ad actually good?

Using the Junbi.ai YouTube creative insights platform, we analysed the AI-Generated Pizza Commercial “Pepperoni Hug Spot.” Unsurprisingly, the analysis revealed that the commercial performed poorly in several key areas. The ad received a low cognitive ease score, which indicates the ad is too visually complex. This shows that the ad has too many visual focal points and rapid scene changes, which makes it difficult for the viewer to process.

Additionally, the brand attention score was also low, indicating that the commercial failed to direct enough attention to the “Pepperoni Hug Spot” brand. However, the ad breakthrough score performed surprisingly well, demonstrating that AI-generated content is successfully able to break through the clutter of YouTube and effectively engage the visual attention of viewers.

Overall, the effectiveness of the AI-Generated Pizza Commercial “Pepperoni Hug Spot” was deemed poor according to Junbi.ai’s analysis.

However, this valuable feedback from Junbi.ai provides advertisers with an opportunity to reassess and optimise their ad content before launching it to a wider audience. By leveraging the insights and recommendations provided by Junbi.ai, advertisers can make informed adjustments in just minutes to enhance the ad’s effectiveness and ensure it resonates more effectively with their target audience.

AI generated beer commercial

Another intriguing example is an AI-generated beer commercial that produces a rather uncomfortable feeling. The ad features people enjoying a neighbourhood BBQ while drinking what seems to be beer from oversized cans and bottles. The AI-generated visuals, set to Smash Mouth’s iconic song “All-Star,” create a peculiar and somewhat unsettling atmosphere. Interestingly, due to the limitations of the technology used, some characters appear with more than ten fingers, adding a quirky twist to the commercial.

The beer advertisement showcases the ability of AI-generated content to push creative limits and provide a remarkable and surprising viewing encounter. Through the use of AI algorithms, marketers can explore unconventional aesthetics and harness the power of unexpectedness, ultimately boosting brand memorability and engagement.

These AI-generated ads demonstrate the potential of AI in revolutionising the advertising landscape. They were produced rapidly, showcasing the time-saving capabilities of AI-generated content creation. But clearly there is still some work to be done!

Best practices for AI-generated advertising

Whilst still being a relatively new field, there are still a few things you can do to stay ahead of the curve with regard to AI-generated advertising. So how can you go about leveraging AI-generated ads yourself? Here are some good tips to get started:

• Stay informed about the latest AI advancements and their role in video advertising. Companies like Adobe and Nvidia will have placed big bets on AI and will be interesting to follow for their AI updates.
• Where available, analyse your data to gain insights into viewer preferences, behaviours, and trends
• Balance AI automation with human creativity to forge emotional connections and align with brand values.
• Invest in ongoing training to enhance skills in AI technologies, data analysis, and consumer behaviour understanding.

Advantages of AI-generated ads

• Ability to keep the ad production in-house, removing dependencies on external production agencies which can be expensive: AI-generated ads provide the advantage of bringing the entire ad production process under the control of the organization, eliminating the need to rely on external production agencies.
• Preventing the sending of generic messages to your entire audience: With AI-generated ads, businesses can leverage their power to personalize their advertising messages.
• Independence from the opinions of creative agencies and production companies: AI-generated ads provide organizations with more control over the creative process. This independence allows businesses to experiment with different ideas and iterate quickly.
• Creation of personalized content for different audiences at the same time, saving time and resources: By leveraging AI-generated advertising, businesses can create multiple variations of ad content targeting different audience segments simultaneously.
• Quick testing and optimization: The AI-generated ads are easy to be tested and provide a great opportunity for quick identifying of the most effective ad elements, and optimizing campaigns accordingly.

Challenges for advertisers

The emergence of AI-generated ads brings forth significant potential impacts on marketers, transforming the way they operate, optimise campaigns, and engage with their target audience. Here are some key aspects to consider:

• Human Oversight and Creative Input: AI lacks the understanding of context, emotions, and cultural nuances that humans possess. Creative agencies and marketers must actively participate to ensure ads align with brand values, resonate with the target audience, and evoke the desired emotional response. For example, an AI-generated ad for a gourmet pizza brand may not capture the nuances and unique selling points that differentiate the brand from standard pizza offerings. Human intervention and expertise are necessary to ensure the effective communication of these complex ideas.
• Limitations of AI to generate life-like content: AI-generated content, while impressive, has some limitations when it comes to creating realistic images. For instance, in the examples discussed above, AI was unable to generate realistic images of people, depicting their hands with more than five fingers.
• Changes in Job Roles and Skills: The integration of AI in advertising will reshape job roles and skill requirements. Marketers must master AI tools, interpret insights, and embrace data-driven strategies. Proficiency in data analysis, AI technologies, and consumer behaviour understanding will be crucial. Moreover, the role of data scientists and AI specialists will expand to optimise AI-powered ad platforms effectively.
• Ethical Considerations: As AI becomes more prevalent in advertising, marketers must address moral concerns. This includes transparency in data collection, privacy protection, and ensuring fair and unbiased practices. Marketers should be mindful of using AI responsibly and balancing personalisation and consumer privacy.

Human oversight and creative input play a vital role in aligning AI-generated ads with brand values, avoiding biases, and effectively conveying complex ideas. By embracing collaboration between AI and human expertise, marketers can harness the power of AI while ensuring impactful and resonant ad campaigns.
So now you can quickly and cost-effectively create AI-generated ads, how are you going to test and optimise all this new content? Junbi.ai is your all-in-one solution for testing your YouTube ad creative, powered by AI so a full analysis only takes a few minutes, and you can test as many ads as you want. Sign up to get a free demo from one of our specialists.