Generative AI is disrupting the entire gamut of content creation and marketing. The content being generated by these AI-based models is compelling and has the potential to give a new flight to human imagination. Still, it is shrouded in controversies that root from universally accepted approaches to sourcing visual content for marketing. For instance, recently, a user of the Chinese social networking app, Xiaohongshu (shau-hong-shoo) has accused the app has used her work to train its AI tool, Trik AI. In this episode, we will explore:
Scientific understanding of why visual content is significant in the human context.
A few real-world examples of marketing campaigns that convey the undisputed place of visual content in marketing.
Emergence of generative AI as a controversial source for sourcing disntinct visuals for marketing.
The likely future of image synthesis and its place in the marketing domain.
Scientific understanding of visual content
Here's a quick question: How much time do you spend on a social media feed? Any social media feed? Nearly every post that you view on your feed has either an image or a video.
Touch is the first sense that a baby develops. As the baby grows, that sense overtaken by visual and auditory senses. No wonder marketers spend significant amount of their time to stimulate those senses to get you to act on their call to action.
In over 100 years, marketers have toyed with imageries in print, voice in radio, and audio-visuals on TV to become visible and appeal to the target audience to purchase their products. Visual medium, today, pretty much dominates any marketers armoury, be it images or videos. Why? Well, here are few interesting scientific findings:
A 2014 MIT study found that people can find images in as fast as 13 milliseconds! Literally, us humans can process an image in the blink of an eye.
When we "see" an image, visual cortex - part of the brain that processes visual input gets activated. An image also activates amygdala - the region that processes emotions and physiological reactions. In nutshell, images impact our emotions faster and more powerfully than words.
People prefer to remember information in a pictorial form compared to in text form. This is known as Picture Superiority Effect. This is because images are more likely to use both visual and verbal channels in brain encoding; and compared to text, images tend to have unique and distinctive visual features. This makes images or videos more memorable.
And finally, depending on the context of the message being communicated, images or videos can help convey trustworthiness of the message being communicated and help build trust in the brand over time.
The undisputed place of visual content in marketing
No one would dispute the leverage images or videos provide marketers to build brand awareness, engage with the target audience, and influence them to buy their products or services. Let's look at a few examples of highly successful marketing campaigns that rocketed the brand presence and sales:
'Think Small' campaign by Volkswagen in 1959 that promoted the benefits of owning a small car, focused on style over function, and challenged pre-conceived notions about what makes a good car. The campaign used reverse psychology to promote Beetle in the US which was different from most cars sold at the time.
'Just Do It' campaign by Nike in 1987 - which encouraged people to pursue their dreams no matter what obstacles they may face.
Then there was the 'Share a Coke' campaign by Coca-Cola in 2011 which used popular names and phrases to strengthen emotional connection between consumers and the brand, and of course, drive the sales up.
Companies have been allocating budgets for procuring compelling audio-visuals to promote their products and services. In the pre-digital era, this meant hiring:
photographers for high impact photos;
illustrators to come up with thought-provoking visuals;
musicians and instrumentalists to compose jingles; and
directors and writers to write quirky scripts for TV commercials.
This created a wide range of sub-industries of creative professionals who came together to develop stories and communicate them in a compelling way to the audience. The advent of digital media brought the emergence of stock photography and video websites like Shutterstock, iStock, Pexels, and Unsplash.This spoilt marketers for choices and achieve scale for their marketing campaigns at a fraction of cost. This also bode well for the creative community as it democratized the discoverability and reach through these websites, and became a reliable source of income. An important thing to note is that both these approaches helped with better management of content licensing which mitigated the legal and financial risks for the brand as well as creative professionals to a large extent.
Generative AI: Emergence of a controversial source for distinct visuals
With the emergence of generative AI, Marketers are now witnessing another source of visuals that is rising like a tsunami over prevailing models of visual content generation. Image synthesis platforms based on generative AI like DALL-E, StableDiffusion, and MidJourney are showing results that are mind-boggling to say the least. Their ability to generate visual content based on text input, refined further with subsequent prompts, is exceptionally good, in high resolution, and near photorealistic.
The Internet is flooded by visually stunning content that is being generated using these AI-based platforms. However, this content is currently wrapped in a legal and ethics debate. And there is, of course, a direct impact on the livelihood of creative professionals.
Generative AI, despite its algorithmic brilliance and unimaginable potential, has become one of the hottest topics in copyright law today. Without getting into the world of complicated legalese, my view is that the expanding space of image synthesis currently rests on three pillars:
Content creators,
End users, and
Generative AI technology.
The AI-powered image synthesis progress is heavily reliant on the endless hours of work put in by true artists and professionals like photographers, illustrators, painters, and musicians. AI models like Stable Diffusion have been trained on massive datasets like LAION-2B and LAION-5B that provide CLIP-filtered image-text pairs. It is apparent that these datasets utilize content created by professionals which is copyrighted and a source of revenue for them. Synthetic images generated based on this dataset neither discloses the underlying images referenced nor provide credit to the original content creator. No wonder that creative professionals are crying foul and have even sued companies behind these platforms.
On the other side are the end users like marketers, designers, social media content creators, and brands that are utilizing the synthesised content for their projects. By and large they are supporters of AI-generated content because a well-framed prompt or a few iterations of the prompt generates content at the speed of light and negligibe cost. It's important to note that companies and brands are still wary of fully bringing such content into their marketing pipeline due to the foggy copyright landscape.
Finally, there is the generative AI technology itself. This technology has been in the works for nearly 5 decades, sitting in the realm of research only. Companies like OpenAI and Stability AI have brought them to the fore thanks to the availability of lightweight training datasets. The flexibility and scalability of their architecture is boosting the accuracy and quality of image synthesis at the speed of light. Technology behemoths like Microsoft, Google and Amazon are investing and supporting such ventures. Not to mention a growing number of VCs and private investors.
Likely future of image synthesis and its place in the marketing domain
Synthetic imagery is one of many areas generative AI has been disrupting over the past year and a half, and that too at a blazing fast pace. It seems unlikely that the development in this space will slow down any time soon.
Does this lead to dooms day prediction to creative professionals and artists? It doesn't seem so. There are rays of hope in the form of EU AI Act or the United States Copyright Office clearly stating that it will not register works generated entirely by AI. Then there are reputed content distribution platforms like Getty Images launching its own generative AI service which has well-defined model for contributor payments to photographers if their work has been used in the training data. Similar announcement has been by Adobe that they are rolling out a payment system for its Firefly system. A few days back, camera manufacturer Leica has announced the Leica M11-P, the world's first camera to create a seamless chain of authenticity from capture to publication.
These developments indicate that their is clear intent to arrive at a sweet spot model that benefits content creators, generative AI services, and partially address the copyright issue, if not entirely.
For companies and brands, embracing synthetic visual content sits in the delta between creating a unique, generative, visual communication for their brand and courage. Courage to take chance on a new technology that can open new areas of garnering the attention of the audience and engaging with them, and proactively supporting the creative community in the event of a dispute.
Comments