What are AI Ads? Artificial Intelligence Generated Ads in Advertising.

Artificial Intelligence is something that was always predicted to be involved in our daily lives at some point in the future. And this future has already happened in some fields, including marketing – with the capability to create both text-based and visual advertising materials for an unlimited number of specifications. However, AI generated ad creation has its own limitations and issues, so we are going to go over this particular topic in this article.

Introduction to AI art generation

Ad creation as a whole is a highly complex process that needs to take a myriad of values into account to be able to produce content that is going to perform as intended in terms of marketing. This is also a highly competitive industry, one where each competitor has to keep up with the current trends to stay relevant and effective with their advertising.

Since a lot of advertisement still relies on regular static images, the introduction of GANs or Generative Adversarial Networks back in 2014 became a gateway for an entirely new chapter of advertising focused on AI image generation. GAN architecture allows for highly detailed images to be generated, with these images being barely distinguishable from the real ones.

This seems like a perfect solution for many advertising companies – exchanging prolonged expensive photoshoots for a technology that is capable of creating an image that fits your very narrow specifications with a fraction of the effort involved.

That’s not to say that AI can only work with visual content – there are also solutions that make artificial intelligence a part of the copywriting team, driven by various advancements in fields such as natural language generation and natural language processing (NLG and NLP, respectively). This makes it even easier to create combinations of text and visuals that are tailored to very specific audience parameters.

The usage of AI in advertising

AI as a whole has been associated with technological advancements for multiple decades now, and it is only natural that different industries would try to implement parts of a new technology to try and get ahead of the competition. The marketing and advertisement industry is no different, allowing for the usage of AI in its work in multiple different ways.

It should not be particularly surprising to learn that AI has been a part of the marketing industry for a while now, with a lot of AI use cases being relatively obscure. This includes chatbots with personalized services, Google’s recommendations as you type, and even custom playlists of music that services like Spotify provide their users on a regular basis.

Of course, this topic cannot be completed without mentioning how AI is used to personalize ads. This includes adapting content for specific users, which includes changing parameters such as layout, color, theme, and even the entire design composition. Additionally, the entire branch of product recommendation ads is AI-driven, as well, including fields such as “similar products based on your preferences” and other variations of a similar message.

At the same time, we have to mention that there is another way to use AI in advertising – to generate images with it for specific promotional purposes. This would be the main topic of this entire article.

AI generated ads

Right now it is possible to see that there are three main AI tools for image generation available on the market – Stable Diffusion, DALL-E, and Midjourney. Marketing agencies are using these tools in a variety of different ways – not only for generating content for advertisement purposes, but also for the sake of brainstorming and saving both money and time on some of the more complicated tasks such as modeling and photo shoots.

All of these tools are relatively new, but they have already managed to open an entirely new branch of marketing, with potential future advantages such as making e-commerce advertising more effective as a whole, as well as a new way to create various special effects or even the ability to generate highly personalized ads for every customer.

Advertisement images generated by AI

Examples of AI generated ads are already around us on many different occasions. Some companies even make the unpredictable part of the AI art generation the main point of their campaigns. For example, Dentsu Creative Portugal recently created an entire campaign of abstract images purely for the sake of promoting Jardim Sonoro – a European electronic music festival.

All of the images in this campaign were created using Midjourney, and the company itself admits to the ad creation process being extremely easy, with them getting the correct images in about three days’ time. The purpose of this campaign was to find something that could be associated with the word “unpredictable” – as the main selling point of the festival, and the use of AI ads turned out to be exactly what they were looking for.

Brainstorming with an AI

It is also fairly common for companies to use AI art generation as a part of the overall brainstorming process. Since the system is capable of generating images based on pretty much any word or sentence out there, it is an invaluable tool for various brainstorming sessions – applying various concepts and ideas to an AI ads generation system to see if it can produce some inspiration for further research and marketing campaign development.

It also proves itself to be more effective than stock photo repositories, especially when it comes to concepts that are far from regular in the first place. Generating AI art is often a lot faster than spending a significant amount of time researching various image hosting platforms for a proper image.

The price difference

Speaking of stock image repositories, the price difference is also not on their side – with examples such as DALL-E offering hundreds of ad generation opportunities for as little as $15 while a single stock image may cost at least $10 or even more. As such, there are circumstances where using AI generated ads is a far less expensive alternative to using traditional stock images and so on.

The future of Artificial Intelligence ads

Even though the technology itself is rather new and a lot of companies are only using it in very specific and rare situations, it is fair to try and see where this entire trend with AI ads going to move. Two of the more exciting options for advertisers that are already talked about right now are custom ads that are truly personalized for every single visitor and the usage of AI to create a multitude of simple special effects.

The former option is a dream of almost every advertising expert, even though right now it is barely a concept in the first place. But there are experts in the field that say it should be possible to create such ads sometime in the future, and this kind of super-focusing on each and every single customer is going to be a great leap in marketing around the planet. For example, if the client in question likes a specific color, the AI in this context should provide advertisement products with this same color in mind.

The latter option is closer to reality in comparison – there are several examples of an AI art being used to generate a multitude of special effects with minimal costs. One of the more notable examples of that is clothing – creating even a dozen of different photo sets with a single person is a logistical and financial nightmare, and AI art is already taking big strides towards being capable of generating similar images while only changing a part of said image, making this concept possible in the future.

However, it’s important to remember that AI tools right now are not the same – Stable Diffusion, Midjourney and DALL-E have their own features and strong points, and the difference is already relatively easy to notice. For example, only Stable Diffusion and Midjourney are capable of creating “not safe for work” images in the first place. DALL-E, owned by OpenAI, does not have that feature and has a rather strict set of filters for what can be generated. Stable Diffusion, on the other hand, is open-source and is also used by some people purely because it somehow creates more natural and nuanced images – which may be a matter of preference, of course.

Potential issues around AI ads

Since AI learns about art in general from preexisting images, there is a good chance of it being taught incorrectly, or not being relevant or balanced in its generation process. Additionally, Artificial Intelligence ads are created after these algorithms are taking images from various image sources, be it Behance, Dribble, Google, etc. As such, there is always a chance that at least a part of the existing image would be recreated by an AI, without giving any credit to the said person whatsoever.

This is where the topic of ownership comes in, which is incredibly complex and nuanced. Since most of these cases would have to be solved on an individual basis, differentiating between an AI creating an image of a specific city or landmark, or an AI being told the name of a digital author and producing that person’s work as a result.

The entirety of this topic becomes even worse when the commercialization of content comes in – using AI to generate a competitor’s product or brand name may very well be outside of what is considered a Fair Use case. As such, some of the bigger platforms are already starting the process of banning AI art from being published completely, with one of the biggest examples being Getty Images.

There are some people that compare this current AI ads generation situation with how it was in the early days of services such as music streaming before there was a proper legal system in place for the author of said music to receive a part of the revenue from his or her music being streamed.

Examples of companies that work with AI in marketing

There are not that many companies that only provide visual AI-generated content for advertisement purposes. However, there are quite a lot of examples when AI in marketing is used as the main selling point for a company’s services. As such, we are going to try and present companies that provide visual content generated by an AI for advertisement purposes.

Synthesia

Synthesia was founded in 2017 by a team of research engineers that specialize in state-of-the-art technologies in the field of computer vision. A lot of the co-founders have earned their names in various research labs and universities around the globe. 

Synthesia provides the capability to create personalized videos with the help of an AI, with two distinct services – Personalize and Translate. The former is a capability to create videos that generate personalized conversations with individual customers, with mentions of their names and business details. The latter is a more well-known service that can translate video content into over a hundred languages with the ability for each generated translation to sound as natural as possible.

DataGrid

DataGrid is a good example of a potential future that we have mentioned above – a variation of the model generation that can create identical models of people with different pieces of clothing. DataGrid uses GANs to create realistic 1024x1024 images of human models for various online clothing stores or advertising campaigns. It has a strong RnD team and the company itself is housed on the campus of Kyoto University in Japan. It was founded by Satoshi Ogawa and Okada Yuki, both of which are majors in ML, or machine learning.

Zalando

Zalando is an online fashion platform located in Berlin that offers a number of cutting-edge marketing solutions to its customers in terms of creative visual advertisements. For example, one of their services is all about generating high-resolution images of models with custom outfits with the ability to transfer both the outfit and the body pose to a different model with ease. 

Zalando puts a lot of effort into researching various improvements to their services with their Zalando Research lab which works with natural language processing, ML, and AI to develop all kinds of solutions for future usage in this sphere.

NEON

STAR Labs is a technological company funded by Samsung that presented its product under the name of NEON back at CES 2020. NEON is an artificial human avatar that claims to be able to both have a conversation with a regular human and sympathize with them, too. This technology is less of a marketing tool and more of a future virtual assistant, and it has two parts to it – SPECTRA and CORE R3.

SPECTRA is the AI engine that is responsible for adding personalities to various virtual personas, and CORE R3 is a component that generates those personalities, including gestures, reactions, expressions, etc.

PixelVibe (formerly Rosebud.ai)

PixelVibe is one more variation of a technology that is capable of generating model faces for specific goals of your marketing campaigns. It claims to be a far cheaper alternative to hiring real models and performing photoshoots. Each model can be adapted for a different target audience, creating near-endless possibilities for advertising. The company claims to have an average of 22% increase in CTR (click-through rate) after the first step of leveraging a marketing campaign with AI generated models included in the package.

It would also be fair to mention that there are plenty of other companies that are not focused primarily on AI in their services but have features that are improved with the help of Artificial Intelligence. Viewst a B2B SaaS platform for automating creative versioning, streamlining approvals and understanding creative performance. Viewst provides AI powered tools for generating copy and analyzing ads designers to run more A/B tests and improve cumulative performance.

AI advertisement and how it must be used

While ad generation with Artificial Intelligence is only starting to become popular in the marketing sphere as a whole, there are quite a lot of opinions already on what it should turn into. A lot of arguments are created towards AI art being more of an additive for the process, rather than being the main tool for content generation. 

AI ads excel in specific areas and tend to be rather ineffective in other situations, so using a combination of the two is a sensible approach to a new technology like this. It is also not a necessity for everyone to start using AI for advertising images specifically, but knowing about these technologies and what they are capable of is always a good idea for anyone knowledgeable enough in the field of digital advertising.

Knowing about the competition in the field is just as important so that the customer knows who to look for when there is a need for a specific service. If the original goal is to get a set of images for a clothing-related company, then Zalando would be one of the first picks. Alternatively, Synthesia would be the first pick of a company that wants to expand its reach in several different languages without paying extra for it. 

FAQ

1. How do automation tools like Marpipe, Smartly, Adcreative.ai, and Viewst specifically streamline the creative process and improve ad campaign performance?

   Automation tools like Marpipe, Smartly, Adcreative.ai, and Viewst streamline the creative process by leveraging AI-driven design systems and data-driven approaches. These tools enable marketers to generate multiple ad variants quickly and efficiently, optimize campaigns based on audience insights, and iterate on the fly. They provide insights into which creative elements resonate most with audiences, enhancing ad performance and improving engagement and conversion rates.

   
   

2. Can these automation tools handle the creation and adaptation of ad creatives for different platforms and audiences?

   Yes, these automation tools are designed to handle the creation and adaptation of ad creatives for different platforms and audiences. They offer dynamic templates, AI optimization, and the ability to generate ad variations tailored to specific target segments. This ensures that the creatives align with the requirements and preferences of various platforms, enabling advertisers to scale their advertising efforts effectively and reach their target audiences with relevant and engaging content. While some of them aremostly work with paid social - Marpipe, Smartly, the others, like AdCreative.ai and Viewst are more universal and can help you with banner ads production for both social and display.

   
   

3. What are the specific features and capabilities of Viewst's automation tools that set it apart from other platforms in the adtech industry?

   Viewst's automation tools offer powerful features and capabilities that differentiate it from other platforms in the adtech industry. Specifically focused on automating and scaling ad creative variations for display ad networks, Viewst empowers advertisers to quickly generate multiple ad variations for efficient A/B testing at scale. Its intuitive interface and automation tools are designed to optimize campaigns with speed and precision. By streamlining the creative process and enabling efficient A/B testing, Viewst helps advertisers make a substantial impact across various ad networks and achieve their campaign objectives effectively.