Beyond the Hype: The Realities of AI in Marketing and Advertising

Seth Geib - Director of Front End Development

As Director of Front End Development at Hart, a full-service digital marketing, advertising and public relations agency, I have witnessed first-hand the effects artificial intelligence (AI), computer systems able to perform tasks that normally require human intelligence, is having on our industry.

To say that AI is changing how we work would be an understatement. In fact, I find myself relying on newly developed tools driven by AI on a daily basis. This change has come rather quickly over the past year or two, with natural language processing and text-to-image tools becoming readily available to the public, and I have reason to believe it will only continue to accelerate at an exponential pace moving forward.

However, it is because of this expedited pace that we must take a step back and think critically about how we are using these tools and the potential impacts they could have, not just on our industry but on our society as a whole. As we navigate how these tools are incorporated and accepted into the world, being open and transparent about their use is important. I am a firm believer that being transparent with our clients about how AI is being used in our work will help set a precedent for AI being viewed as an additional marketing tool, and not seen as a cheat or “easy button.”

AI has the power to increase the output of nearly every aspect of the advertising industry, from deciphering analytics and making predictions for user behavior, to writing content and copy, to generating imagery and video. We’ll be taking a look at some of the tools on the market right now, reviewing how they can be used and pondering the implications for our industry.

Where Did This Come From?

It’s a fair question – to most, the rise of these tools powered by AI algorithms must have seemingly come out of nowhere. AI’s meteoric ascent can largely be attributed to the release of OpenAI’s ChatGPT model on Nov. 30, 2022. Within less than a week, ChatGPT amassed 1 million registered users, and by January 2023, that number grew to 100 million, making it the fastest-growing user base in software history.

Behind this global surge is ChatGPT’s myriad of uses, from writing email responses and video scripts to programming code and jokes (with subjective success) all with minimal human intervention. The success of ChatGPT has even influenced industry giants like Microsoft to look toward AI solutions – the company is now working to incorporate its own AI chatbot into its Microsoft Office 365 product and Windows operating system.

Beyond the Written Word: Image-Generating AI

Another significant development in the artificial intelligence landscape has been the emergence of image-generating tools. In the same time period, Stability AI released its open-source image-generation tool, Stable Diffusion, to the public, which enables users to generate completely unique images from simple text descriptions, even without internet access. Other image-generation tools like Midjourney and DALL-E, as well as major players in the graphic design industry, have all hit the market en masse.

For example, Adobe’s AI Firefly software is now being integrated into the latest beta versions of Photoshop, allowing users to remove or replace elements in an image, extend the image resolution and even create entire images seemingly from thin air. Adobe also plans to incorporate Firefly into video editing tools like Premiere in the near future, so users can create entire video concepts from a text prompt.

All AI content data repositories are sourced from human-generated content, meaning many artists’ content has been used to train models like Stable Diffusion and Midjourney, leading to a number of copyright infringement suits. Adobe is working to be transparent with its data sourcing and claims its Firefly model is being trained solely on Adobe-owned or copyright free content.

Application in the Workplace: What to Keep in Mind

Artificial intelligence is still being used for more “traditional” machine learning use-cases as well, like analyzing big data and user analytics, but the advent of natural language models like ChatGPT has made the output of these applications much more user-friendly. In fact, the latest beta of ChatGPT’s plugins allows users to upload and analyze their own files (PDFs, XLSX, CSV, etc.) and Python code, all with a natural language response.

Additional AI services like Neurons are being used to analyze creative concepts, allowing agencies to test their creative campaigns in an AI-driven testing environment and make adjustments before releasing them to market. With any AI tool, however, the algorithms and even the data that feeds these predictions can have their own inherent biases, such as user gender, race, age and the like. To use these new tools in our work, we all must keep this in mind when interpreting results and when providing data sets to these AI models. The more we can learn about AI tools and the chances of these biases occurring, the more we can work to avoid them in our deliverables.

It's also important to note that while AI can enhance creative work and design, it’s not a replacement for human expertise. Agencies still need to use their creativity, empathy, relationships with clients and their own extensive experience to create products and services that meet the nuanced needs of their clients. AI is just another tool in the toolkit.

With the use of these AI chat models becoming so prevalent, OpenAI developed a classifier that can parse content and determine if it was generated by an AI. Because content can always be checked for AI involvement, our industry should rely on transparency with clients so they feel more comfortable with AI being used as a tool in creative marketing. Moreover, it should be thoroughly explained to clients this is a tool that can improve and expedite deliverables, not replace the hard work we are doing for them.

Here at Hart, we are working to use AI in the most responsible and transparent way possible, though like the rest of you, we are learning as we go, and it is changing every day. While that can be a difficult thing to manage, I believe this is an exciting time for the evolution of our industry. As long as we meet these challenges with curiosity, transparency and grace, the continued integration of AI tools into our everyday lives can be something we embrace, not cast aside out of fear and lack of understanding.

If you’re interested in learning more about AI in the marketing industry, let’s talk.

Blog | AI