Andrew is VP of Marketing Strategy at VALTIM, where he helps cause-based organizations improve member/donor communication channels.
Getty
In 1996, Bill Gates wrote an essay entitled “Content Is King.” This missive highlighted the importance that content would play as the internet took off, particularly as it related to driving advertising dollars to content creators, whose content would otherwise only be monetizable through a “fairly high subscription rate.” Gates would go on to write, “Those who succeed will propel the Internet forward as a marketplace of ideas, experiences, and products—a marketplace of content.”
Gates’s vision, cast literally at the dawn of the internet, would hold true. Several years later, Google would launch its search algorithm, scraping up content from across the internet and dishing it out to anyone who searched using keywords, keyword phrases and, most recently, natural language queries.
Today, businesses and marketers everywhere invest millions of dollars in generating content designed to appease these enigmatic search engine and social media algorithms to get their content—and their brands—to the top of consumers’ screens.
Content Marketing In Practice
At its core, content marketing serves to reinforce the theory of the Hierarchy of Effects within the consumer’s purchasing journey—first, in the initial discovery/awareness phase and also to provide product knowledge, education, building brand affinity and, ultimately, the conative conviction that the product/brand meets the consumer’s needs, driving them to conversion.
Until recently, most content has involved painstaking research, deep subject matter expertise, hours in front of screens writing and rewriting scripts/text, and—depending on the type of content—organizing teams of specialists to build out multimedia assets associated with the content. However, with the advent of machine learning, several tools have recently captivated the attention of marketing professionals and the general public alike—particularly the ease with which content can be created in a matter of seconds rather than days. These tools market themselves as artificial intelligence (AI) systems but are, in reality, little more than sophisticated natural language models that are able to interpret simple inputs and provide contextually relevant responses based on information that is made available to that model.
These AI tools—most notably GPT (the model underlying OpenAI’s ChatGPT software)—offer a potentially game-changing and disruptive model to content marketing. Marketers simply need to feed a piece of SaaS software a few keywords, and instantaneously have an output that targets their specific content needs. Songwriters, artists, copywriters, programmers and a whole host of other creatives who have spent years learning, practicing and perfecting their trades are seeing a piece of code provide—at least on the surface level—reasonably convincing parallels to their own work. The speed to market is unrivaled and certainly concerning.
Natural Language Model’s Weaknesses
An analysis of the outputs of these natural language models (not to be confused with more analytical data-driven models) reveals some interesting weaknesses of these tools. Chief among them is that its text results—no matter how carefully the prompt is written—fundamentally lack the authority of human-created content. Put another way, natural language model output is reliant on its access to preexisting content that it has been specifically trained to recognize and interpret.
At present, and certainly subject to change, these tools mimic the content they have access to. The models currently available to the market are incapable of comprehending the depth of their subject in a way that enables the software to structure responses in a manner that can provide an authoritative response on the subject matter.
For instance, I provided the ChatGPT model (via a third-party SaaS interface) with the following prompt: “Write a blog post on the benefits of leveraging direct mail in an omnichannel world.” Here is an example output from one of the paragraphs:
“When used effectively in combination with other channels, direct mail can help companies increase brand awareness and reach potential customers who may not be using the other channels. Additionally, it helps create a more personalized experience for customers by providing them with tailored offers that are relevant to their interests and needs. This allows companies to stand out from their competitors by delivering messages that are specific to each customer’s individual situation.”
While broadly accurate, the model’s output reflects only a freshman marketing student’s understanding of its requested subject matter, failing to provide specific examples or context for its declaration. In similar tests with other queries, its responses are consistently vague and provide only the highest level of understanding of what it is specifically being asked to produce.
Even when presented with leading questions to build an article piece by piece, it’s apparent that these tools lack the ability to make inferences and comparisons, provide firsthand experience and anecdotes, or complete a truly deep authoritative dive into their content. Instead, it is dependent on the authority, reliability and originality of the content creators whose work it imitates. Thus, the success and future of AI content creation are not at the expense of original content creators but rather at those content creators’ mercy.
Content Kings
In the end, any machine-learning algorithm is dependent on the information that it’s been provided to model its responses. AI, therefore, takes on its most accurate definition: imitation of intelligence. It can go through the iterative steps necessary to complete the task it’s been requested to take. However, the quality of its output can never exceed its input. At least for the near-term and mid-term future, natural language and “creative” AI-based models will heavily depend on the authority of their source content when serving their responses.
The most valuable source content will come from those marketers who are able to establish themselves as authorities within their subject matter expertise. In this case, content quality will become more important than content quantity. These “content kings” will be leading the charge in the content marketing efforts moving forward.
Forbes Communications Council is an invitation-only community for executives in successful public relations, media strategy, creative and advertising agencies. Do I qualify?
