Getting Creative with AI

2023 is the year when artificial intelligence transformed from science fiction to workplace reality. One million people used ChatGPT in its first five days, and within three months of launch it had more than 100 million users. If you work in employer branding, marketing, or communications, you’ve probably spent time this year reading about, thinking about, and possibly worrying about AI.

The hype has died down—according to Google Trends, searches on “ChatGPT” and “generative AI” peaked at the end of April but have declined by 50% since. We are now in the stage when AI is less something that people are talking about, and more something that they are trying to use: reportedly, 61.4% of marketers have used AI in their jobs, and 44.4% have used generative technologies (that is, AI systems that can generate text or images) for content production.

These trends clearly suggest that AI is here to stay, and that proficiency with these tools could well become an important job skill going forward. Among those who have not started using these new technologies, the number one reason cited is a lack of understanding. This is a rapidly developing landscape, and it can be hard to get your bearings, but we’re here to help. Read on for a rundown of how, specifically, generative AI can be an asset in the creative aspects of employer brand marketing.

 

Getting the Partnership Started

If your employer has not yet provided an approved tool, you can build familiarity with generative AI’s capabilities by creating an account with OpenAI or running searches on Microsoft’s Bing or Google’s Bard. These tools are built around a conversational interface: simply type in a question or make a request, and then see how the AI responds. When using generative AI, it’s important to keep in mind that you, the human, have an important role to play. Large Language Models (LLMs) like ChatGPT are infamous for “hallucinations”: every now and then they are confidently wrong about the facts. This is not always the case, but it happens often enough that you need to fact-check anything the AI produces that you don’t already know to be true. 

AI can be a valuable partner in your creative work, but—as is the case with self-driving cars—keep your hand on the wheel. Rather than simply handing off your work to the machine and hoping for the best, you should work in partnership with the AI: Take what ChatGPT gives you and then tweak, revise, and upgrade it. In many ways, these new tools turn writers into re-writers; the good news is that your partnership with AI allows you to work faster and get more done while keeping the best parts for yourself.

 

It's All About the Prompt

Working with AI begins with your prompt: the question or request that you input into the system. A poorly constructed, vague, and unfocused prompt will produce disappointing results. A well-constructed prompt, by contrast, is quite specific and includes five elements:

1. The role that the AI should play (for instance, you might type into the chat window: “You are an employer brand manager writing an article for a careers blog”).

2. Background details that bring the desired output into sharper focus (the second sentence of your prompt might read: “Your goal is to interest software engineers in career prospects at your company”).

3. The specific objective that the AI should aim for (“You are writing an article on how blockchain technologies are an exciting area of development within the telecommunications sector”).

4. Boundaries or conditions that might limit the AI’s response (“Your article needs to be between 700 and 1000 words in length.”)

5. Any additional instructions to further refine the output (“The tone should be light and conversational”).

Will a quality prompt produce output that you can simply copy-and-paste into your company’s CMS? Almost certainly not; for one thing, the AI will not have your level of insight into your company, your brand, or your workforce. You will, however, be off to a running start and well on your way to a quality piece of work, and you can partner with the AI to add these elements and make the article great.

 

Cure Your Writer’s Block

Everyone who has written on deadline understands how daunting a blank page can be. Sometimes the hardest part of writing is getting that first word on the page. Generative AI can be a big help, because the machines don’t share our anxieties. No matter what you ask, the AI will produce it almost immediately.

This quality makes generative AI a valuable (and tireless) idea machine. Are you drawing a blank on a social media post on your company’s commitment to work-life balance? Let the AI take a first crack at that. An example prompt could be: 

“Act as a social media manager. You are writing a Twitter post on the topic of how important it is to take your PTO. The tone should be upbeat, and should celebrate employees who are taking time with their families and for themselves. Provide ten variations on the tweet.”

Now, instead of a blank page, you have ten different possibilities. None of them will be perfect, but take what works, piece those parts together, and then add a little twist of your own—another great example of partnership in practice.

 

Try On Different Styles

AI can also be helpful in working with tone and style. Maybe there’s a brand voice that doesn’t come naturally to you, or perhaps your default style is a little too serious (or not serious enough) for what you’re working on. AI can help here, too. 

In this arena, you get the best results when you’re able to train the AI on previous examples of the content you’re aiming for. One service (currently in beta, so free for now) is Lex, which includes a feature for AI to continue writing from the point you left off. It bases its suggestions on the content already on the page, so you can fill a document with examples of what you’re aiming for, and then Lex’s suggestions will seek to match the tone and content of these examples.

You can go farther, though. Many paid AI services allow you to define the tone for a piece of writing, but even ChatGPT will rewrite a passage along the guidelines you include in your prompt. You can even ask it to rewrite your passage in the style of someone famous—Malcolm Gladwell if you’re going for clear and concise, or maybe Barack Obama if you’re reaching for some eloquence in your copy. As always, this should not be your final step; consider the AI’s suggestions but then make them your own.

 

What About Imagery?

By this point, you may be thinking, “Can’t AI generate images for my articles and social posts?” The answer is “yes and no.” There are powerful generative systems designed to produce images. Two of the most famous are Midjourney and DALL-E, but these capabilities are now also showing up in widely-used tools such as Photoshop. The aptitude of these tools is astonishing, but it also raises important questions. For example, Midjourney is currently facing a copyright infringement lawsuit, based on the fact that it was trained on images that were under copyright.

Adobe is not (yet) facing similar scrutiny, but the ability of Photoshop Generative Fill to manipulate photographs is troubling. If you have a great employee photo in a call center, but you want to show them in a retail location, you could use Photoshop to change the background or even the uniform they’re wearing—but should you? 

A photograph of an employee in their own environment conveys authenticity; as soon as you fundamentally change that photograph, it becomes inauthentic. That can be dangerous for brands, because once employees and candidates perceive them as inauthentic, it will be very hard to get that trust back again. Now is a great time to kick off conversations–if they’re not already happening–on these new tools, what your brand standards and guidelines are on photo manipulation, and what techniques fall short of your values.

 

Use AI with Authenticity

The primary risk of generative AI may be that it will lead to the proliferation of bland, generic, inauthentic brand voices. Author and AI thought leader Nina Schick predicts that, by 2025, AI systems might produce 90% of online content. The risk of inauthenticity, though, doesn’t mean you shouldn’t use generative AI; it simply means that you should use it to improve your writing rather than simply outsourcing that work to a machine. Whether it’s an email, a blog, a tweet, or a job post, you have something special to add that a machine can never simulate.

At the end of the day, we are all in the business of creating and sustaining relationships with candidates, employees, social followers, and visitors to our websites. Authenticity is a critical element of any lasting relationship. AI used poorly works to the detriment of authenticity, but used strategically, it can accelerate your work and free up time you can invest in creating something truly human, authentic, and captivating to your audiences.

Related Posts