The buzz around ChatGPT (Generative Pre-trained Transformer) took flight on February 7, 2023, when OpenAI — an artificial intelligence research and deployment company — asked a limited number of people to test it. The initial reviews were rave, and by the end of March millions of people worldwide were using ChatGPT to create content. A student in the UK asked ChatGPT to write a letter to the city council protesting a parking ticket, and her fine was revoked. And this is just the beginning.

For the uninitiated, if there are any of you left out there, ChatGPT is one of the most popular and widely used large language models (LLMs) in circulation. ChatGPT can follow complex instructions given in spoken language and solve challenging problems accurately — imagine if Alexa or Siri could contribute to a creative process. Its capabilities are as astonishing as they are simple:

  • generate, edit, and revise in collaboration with users on creative/technical writing tasks, including songwriting and screenwriting
  • learn an individual user’s writing style
  • accept images for inputs and generate captions, classifications, and analyses
  • handle more than 25K words and support longform content creation, extended conversations, and document search/analysis.

Google, Facebook, Microsoft, and other tech companies have developed similar models. And AI is stretching far beyond just smart chatbots. There are already several iterations of:

  • text-to-image models that create realistic images from natural language
  • image captioning models that describe pictures with words
  • open-source applications that will classify and summarize text
  • vision libraries that enable computers to detect and track objects.

These tools have spurred a tremendous amount of both excitement and distress in the world of work. Will AI take our jobs? Make us more productive? Or a combination of both?

Let’s examine ChatGPT as it’s used in the advertising industry specifically for a closer look.

The Limitations of AI

From OpenAI, the developer of ChatGPT: “While we have safeguards in place, the system may occasionally generate incorrect or misleading information and produce offensive or biased content. It is not intended to give advice.” OpenAI admits to three core limitations of ChatGPT:

  • may occasionally generate incorrect information
  • may occasionally produce harmful instructions or biased content
  • limited knowledge of world and events after 2021.

These limitations bespeak the need for human participation in any and every AI project. According to Todd Reinhart and Bill Skrief of AdAge, AI should streamline the human creative process, but it shouldn’t produce final deliverables:

“Just because a technology is in use doesn’t mean it’s ready for prime time. AI has been a media darling, but creative leaders need to look beyond the hype to determine if AI is right for their process and internal needs and if it synergizes with the comfort level and requirements of clients and partners.”

Reinhart and Skrief warn of the necessity to properly govern AI. Essentially, users must be fully aware and transparent about what it can and can’t do, as well as what it should and shouldn’t do.

For example, FreedomGPT is an LLM-based chatbot trained to have neither guardrails nor inhibitions. The tool was built by AI venture capital firm Age of AI and dispenses with most forms of censorship: it can explain how to build a bomb, it can use racial epithets and slurs and will, with the right input, happily praise Adolf Hitler. This has some unsettling implications. As recently as April, President Joe Biden told his council of science and technology advisors it remains to be seen if artificial intelligence is dangerous, but technology companies must ensure their products are safe.

Right now, ChatGPT is going through growing pains and working through a number of ethical and legal issues. It’s become apparent that human oversight is still required to recognize the technology’s potential for bias and toxicity.

How AI Can Help Ad Agencies

With only a few key phrases, ChatGPT can quickly generate multiple ad copy options. This can shorten the time usually devoted to brainstorming sessions or first drafts, giving copywriters and editors a head start. How simple is it? Ask ChatGPT, “write me ad copy for ______________.” So let it be written, so let it be done.

ChatGPT can also analyze consumer data and identify the most effective ways to reach specific consumer groups. The program can:

  • help advertisers determine the best social media platforms to use for their ad campaigns
  • determine the most effective times of day to target a particular market
  • generate new interests and audience groups.

ChatGPT can also conduct market research, another boon to ad agencies. This is immeasurably useful for advertisers looking to target consumer groups and/or specific demographics and discern their likes, their interests, and their overall buying personas. Some advertising experts even predict ChatGPT will eventually provide wireframes for the development and deployment of individual campaigns.

How AI Can’t Help Ad Agencies

AI is a good tool for generating new ideas and first drafts, but from there, think of it as a cake pulled from the oven: you still need a pâtissier to ice and slice it and put it in a display case (which said pâtissier has set at the perfect temperature and arranged alluringly, etc.). An application like ChatGPT can give you a good starting point, but it takes human creativity, experience, and imagination to create something wholly memorable.

“…at best, ChatGPT is a tool, a powerful one, that can save time and help creative talent focus on bigger thinking, but at worst, it’s a shortcut to mediocrity at scale,” says M.T. Fletcher of AdAge. “You don’t need a gizmo; you need an idea. You need an idea for a story, because stories are what people remember. Make your consumer the protagonist, your brand the catalyst, and tell a tale of transformation, redemption or discovery. From myths to memes, stories are what we share. And campaigns are never-ending stories. Which means creativity and non-linear thinking are still the most powerful tools in your arsenal.”

Example: AI in Action

As an example of ChatGPT’s capabilities and limitations, we asked it the question in the title of this article: “AI and Advertising: How Should Agencies Balance Smart Tech and Human Talent?” Here’s its verbatim answer:

“Artificial intelligence (AI) has transformed the advertising industry, enabling advertisers to target specific audiences, optimize ad campaigns, and measure results with greater accuracy. However, despite the benefits of AI, it is important for advertising agencies to strike a balance between smart tech and human talent.

 Here are some ways that agencies can balance AI and human talent in advertising:

  1. Identify areas where AI can add value: Agencies should identify areas where AI can add value, such as targeting specific audiences, analyzing data, and optimizing ad campaigns. This will enable agencies to leverage AI to improve their advertising efforts while also allowing human talent to focus on more strategic and creative tasks.
  2. Hire talent with the right skills: Agencies should hire talent with the right skills to work alongside AI technology. This includes individuals who have a deep understanding of AI, data analysis, and technology, as well as those with a strong creative skill set who can think outside the box.
  3. Encourage collaboration: To strike a balance between AI and human talent, agencies should encourage collaboration between the two groups. This will enable team members to share knowledge and expertise, and to leverage each other’s strengths.
  4. Emphasize the importance of creativity: While AI can help optimize ad campaigns, it cannot replace human creativity. Agencies should emphasize the importance of creativity in advertising and encourage their teams to think creatively and come up with innovative ideas that resonate with audiences.
  5. Use data to inform decisions: AI can help agencies collect and analyze data, but it is important to remember that data is only one piece of the puzzle. Agencies should use data to inform their decisions, but also rely on the insights and expertise of human talent to make informed decisions.

In summary, advertising agencies should strike a balance between AI and human talent by identifying areas where AI can add value, hiring talent with the right skills, encouraging collaboration, emphasizing the importance of creativity, and using data to inform decisions. By doing so, agencies can leverage the benefits of AI while also leveraging the unique strengths of human talent.”

As you can see, it’s a great start, but it’s largely just a start. The article is grammatically sharp with solid pacing, syntax, and tone, but the content overall is very general and offers few unique insights that truly educate the reader.

How the Right People Can Put AI to Work for Your Agency

AI content will often be “pretty good,” and “pretty good” is impressive when produced with such ease and speed. But anyone in advertising, communications or marketing knows that “pretty good” isn’t enough, especially in an increasingly incredible competitive market during economic headwinds.

So will the phrase “flesh-and-blood” become the “bricks-and-mortar” of the ‘20s, a way to describe increasingly irrelevant human capital as AI continues producing more and better content? In short, no, it shouldn’t.

Replacing your copywriters, editors, and proofreaders with large language models, or your animators, graphic designers, and illustrators with text-to-image AI applications, is a recipe for disappointment and failure. If you put together an orchestra without a conductor or set sail without a captain, you’re sure to find yourself face to face with disaster. In other words, you can’t just leave AI to AI.

AI is likely to change how your agency works and how your talent spends their time. It can grant you efficiency and increased speed by enabling you to produce more content in shorter periods of time. But let’s not get ahead of ourselves — AI is not likely to outright replace employees en masse anytime soon. At least, not for smart agencies and forward-thinking talent.

Because AI is so accessible and affordable, it’s helping to establish a baseline of competency that every agency can easily reach. To stay competitive, it’s more important than ever to have the very best creative talent. Brands will need the writers, editors, project managers, analysts, and more who can write careful prompts, fact-check and edit AI output, fine-tune content for specific customers, propose novel approaches to common client challenges, delight audiences with original ideas and, perhaps most importantly, stay keenly aware of AI’s quickly evolving abilities and pitfalls.

Jobseekers who stubbornly stick to old methods may soon find themselves becoming more and more replaceable. But the open-minded creatives who are eager to learn new tools, adapt their processes, and consistently challenge conventional thinking are going to be the most impactful contributors that a company can have. This is the type of talent that successful agencies will be seeking out, developing, learning from, and utilizing in the years ahead, and it’s the type of talent we work with at Creative Circle.