The rapid integration of generative artificial intelligence into the professional communications landscape has presented a paradoxical challenge: while these tools offer unprecedented efficiency, they simultaneously threaten to erode the distinctiveness that defines successful brands. As organizations increasingly rely on Large Language Models (LLMs) to draft press releases, social media copy, and internal communications, the industry is witnessing a phenomenon often described as the "sea of sameness." This trend toward homogenized content has prompted experts to call for new standards in how human creativity and machine efficiency coexist.
Karen Freberg, a professor of strategic communications at the University of Louisville and a prominent voice in the digital marketing space, recently addressed these concerns during Ragan’s Writing Certificate course. Her central thesis posits that the ubiquity of AI tools necessitates a more rigorous approach to brand identity. "If we all are using AI, we’re all sounding like each other," Freberg noted, highlighting the inherent risk of using tools that are trained on vast, generalized datasets to produce specific, branded content. Rather than advocating for the abandonment of AI, Freberg suggests that communicators must evolve their methodologies to ensure their unique organizational "soul" remains intact throughout the digital drafting process.
The Evolution of Automated Writing: A Brief Chronology
To understand the current crisis of voice, it is essential to trace the trajectory of AI in the writing profession. The journey from basic corrective tools to generative powerhouses has been swift, fundamentally altering the workflow of PR professionals and journalists alike.
- 2010–2018: The Era of Rule-Based Assistance. During this period, tools like Grammarly and early versions of Microsoft Word’s editor focused on syntax, spelling, and basic style rules. These tools were reactive, correcting human errors rather than generating original thought.
- 2019–2021: The Emergence of Transformer Models. With the release of OpenAI’s GPT-2 and later GPT-3, the industry saw the first glimpse of "predictive text" on a massive scale. While impressive, the output often required heavy human intervention to maintain factual accuracy and tone.
- 2022–Present: The Generative Explosion. The launch of ChatGPT in November 2022 marked a turning point. For the first time, high-quality, long-form content could be generated in seconds. This led to an immediate adoption surge across marketing and communications departments globally.
As of 2024, industry data suggests that over 75% of communications professionals use AI in some capacity for content creation. However, a 2023 study by Salesforce indicated that while 60% of marketers use generative AI, nearly half are concerned about its impact on brand consistency and the potential for "hallucinations" or factual errors.
The Concept of the "Voice Bank"
One of the most significant pitfalls identified by Freberg is the inadequacy of traditional brand guides when applied to AI prompting. A standard brand guide might describe an organization’s voice as "friendly," "bold," or "professional." While these adjectives are useful for human writers who can interpret nuance and context, they provide insufficient direction for an AI.
To combat this, Freberg recommends the development of a "voice bank." This is a specialized repository of linguistic assets designed specifically to "train" or prompt AI models. A comprehensive voice bank includes:
- Signature Phrases: Specific idioms or slogans that the brand uses consistently to build recognition.
- The "Never" List: A documented list of words and phrases that are strictly prohibited because they conflict with the brand’s identity.
- Syntactic Structures: Observations on sentence length and complexity. For example, a tech startup might favor short, punchy, action-oriented sentences, while a law firm might require complex, subordinate-clause-heavy structures to convey authority.
- Emotional Anchors: Specific adjectives that go beyond the surface level, defining how the brand should make the reader feel.
By feeding these specific parameters into an AI prompt, communicators can move beyond the "average" output of the model and force it to operate within the specific constraints of the brand’s personality.
Case Study in Contrast: Liquid Death vs. Traditional Wellness
The importance of a defined voice bank is perhaps best illustrated by the beverage brand Liquid Death. In a market traditionally dominated by "soft wellness" language—using terms like "refreshing," "pure," "hydrating," and "tranquil"—Liquid Death opted for a radical departure. Their branding is aggressive, irreverent, and borrows heavily from heavy metal and punk rock aesthetics.
Freberg points out that if a communicator at Liquid Death used a standard AI prompt to write a product description for water, the AI would likely default to the "fluffy" language typical of the industry. "You never have the soft wellness, very fluffy language," Freberg said of the brand. "Absolutely not. That is not Liquid Death."
For Liquid Death, a voice bank would explicitly forbid words like "serenity" and instead prioritize "murdering your thirst." This level of specificity is what prevents the AI from diluting a high-equity brand voice into a generic marketing script.
Supporting Data: The Impact of Brand Consistency
The push for unique voice is not merely an aesthetic preference; it has tangible economic implications. According to data from Lucidpress, consistent brand presentation across all platforms can increase revenue by up to 23%. Conversely, inconsistent messaging—which is a high risk when using unrefined AI—can lead to consumer confusion and a loss of trust.
Furthermore, a 2024 report by Edelman on Brand Trust revealed that 63% of consumers are more likely to buy from a brand that "feels like it has a soul" and demonstrates a clear, human-centric personality. As the market becomes saturated with AI-generated content, the value of "human-sounding" or uniquely branded content is expected to rise, creating a competitive advantage for those who master the art of the AI-human hybrid workflow.
Strategic Implementation for Communications Teams
For organizations looking to protect their voice while leveraging AI, industry analysts suggest a multi-tiered approach:
1. The Human-in-the-Loop Protocol
No AI-generated text should move directly to publication without a human "voice editor." This role is distinct from a traditional copy editor; while a copy editor looks for grammar and facts, the voice editor ensures the piece aligns with the brand’s specific "voice bank" and emotional resonance.
2. Custom GPTs and Private Models
Many enterprises are now moving away from public versions of ChatGPT in favor of private, "fine-tuned" models. By training an AI model on an organization’s historical archives—years of successful press releases, speeches, and internal memos—the model begins to "learn" the specific cadence of that organization, reducing the need for extensive manual prompting.
3. Prompt Engineering as a Core Competency
Communications teams must view prompt engineering as a foundational skill. A prompt should not be "Write a blog post about our new product." Instead, it should be: "Write a 500-word blog post about our new product using the following voice bank parameters: use short, punchy sentences; avoid all wellness-related adjectives; include the phrase ‘disrupting the status quo’; and maintain a tone of irreverent humor."
Broader Implications and the Future of the Profession
The shift toward AI-assisted writing is forcing a re-evaluation of the communicator’s value proposition. If the machine can handle the "what" (the facts and basic structure), the human must handle the "how" (the tone, the ethics, and the strategic alignment).
There is also a growing discussion regarding the SEO implications of AI homogenization. Search engines like Google have updated their algorithms to prioritize "Experience, Expertise, Authoritativeness, and Trustworthiness" (E-E-A-T). Content that sounds like a generic AI output—lacking unique perspective or a distinct brand voice—is increasingly penalized in search rankings. Therefore, maintaining a unique voice is now a requirement for digital visibility.
In conclusion, the rise of AI does not spell the end of the unique brand voice; rather, it raises the stakes for its management. As Karen Freberg and other industry leaders suggest, the path forward involves a disciplined integration of AI tools, underpinned by a deep, data-driven understanding of what makes a brand’s communication unique. By moving from broad descriptors to granular "voice banks," and from passive users to active prompt engineers, communicators can ensure that their organizations continue to stand out in an increasingly automated world. The goal is not to sound like a machine, but to use the machine to sound more like a refined version of the brand itself.
