At Flow Communications, like many others, we’re deeply engaged in the ongoing ethical debate surrounding the integration of artificial intelligence (AI) in our creative processes, particularly in crafting compelling campaigns for our clients.

In the past year, as AI technology has become increasingly prevalent, we’ve grappled with numerous questions, particularly regarding content generation: Is it ethical to employ AI to generate copy for our clients? Should we even consider it? If we do decide to utilize AI, what are the best practices? How transparent should we be with our clients about our use of AI, if at all? And why should clients pay for AI-generated content when they could potentially do it themselves at no cost?

These are complex and multifaceted questions that require careful consideration of both ethical principles and practical implications.

While we’re not opposed to AI technology like ChatGPT—in fact, we find it incredibly exciting—we recognize its strengths lie in specific tasks, such as summarizing lengthy text, generating listicles, brainstorming ideas, or conducting limited research. However, when it comes to creative copywriting, AI falls short. It’s not so much a writer as it is a tool for writers, akin to a digital dictionary or thesaurus.

In essence, AI lacks the capacity to truly ‘write’ in a creative sense. Its output tends to be bland, unimaginative, and often repetitive, lacking the human touch that makes for engaging content. For instance, phrases like “unwavering commitment” paired with “resilience” and “support” are common giveaways of AI-generated text, often presented in predictable formats such as lists of three with the Oxford comma. Additionally, AI-generated sentences may sound sophisticated but lack genuine meaning or impact.

During a recent internal training session, we examined copy created by ChatGPT for a hypothetical invitation. In just two paragraphs, we identified over 15 verbal, punctuation, and syntactical cues that revealed its non-human origin.

It’s becoming increasingly clear to our clients as well. We’ve had instances where clients have attempted to use AI to generate copy themselves, only to find it lacking and request our assistance in refining it. While AI technology will undoubtedly improve over time, it’s essential to recognize that it can’t replace human writers—at least not yet.

AI may simulate certain aspects of human intelligence, but it remains just that: a simulation.

According to Flow Communications’ chief technology officer, Richard Frank, AI essentially functions as an advanced autocorrect system, adept at mimicking human behavior—especially when it’s been inadequately trained.

While AI-generated content has become a mainstay, it’s important to recognize its inherent flaws. Despite its speed in producing content, the inevitable need for extensive editing, fact-checking, and humanization often outweighs any initial time savings. In fact, considering the amount of time required to refine AI-generated copy, one might as well write it from scratch.

Editing is particularly crucial because AI-generated content encompasses a wide array of material, including biases, falsehoods, and potentially harmful content like misinformation, conspiracy theories, misogyny, and racism. Without the ability to discern between right and wrong, AI poses significant risks, such as inadvertently creating campaigns that perpetuate racism or sexism or deny historical atrocities like the Holocaust.

Furthermore, it’s essential to acknowledge that AI is inherently flawed and can be deceptive. It operates as a digital “thief” and “con artist,” implicating users as accessories after the fact in its production of content.

Let’s address one of the age-old issues in literature: plagiarism. AI, in its rapid assembly of content, often mimics, borrows, or outright appropriates others’ intellectual property without proper attribution. While it may not always fit the strict definition of plagiarism, it certainly resembles it, as it involves presenting someone else’s work as your own.

Now, consider the scenario where you utilize AI-generated content for a client, expecting payment in return. While it might not technically constitute fraud—defined as criminal deception for financial or personal gain—it certainly bears resemblance to it. It involves being less than forthcoming about the origin of the creative work, all while seeking compensation for it. Would you feel comfortable with this arrangement if you were the client?

Let’s revisit the questions we continuously ask ourselves:

Can we use AI for clients? Technically, yes, it’s possible.

Should we? Absolutely not. AI lacks the inherent creativity, intuition, and human touch required for effective copywriting.

What should we communicate to clients regarding AI usage? The truth. It’s imperative to maintain honesty and transparency by disclosing whenever AI is employed in the creative process.