Ethical AI and Writing: Can the Two Co-Exist?
By LeAnne DeFrancesco, vice president at Vanguard Communications
If you make a living from writing, your hackles have likely been up for a while about artificial intelligence (AI)--not the customer service chatbots on a website “Can I help you find what you’re looking for?” or even the expanding search capabilities. I’m talking about generative AI that puts together words and phrases that make sense, are creative and do the legwork of writing in an instant.
Wait – Was that my career that just flew out the window?
Writers, be comforted. Your career is intact, but there’s no reason you can’t make AI your bestie. In fact, we know that writers and authors are using AI as a creative and organizing tool, not a replacement for the beautiful subtleties and nuances of language and persuasion that only humans can provide.
And that’s really where the intersection of ethical AI and writing lies: Can we find a way to use AI to help us do our jobs better without passing it off as the work product of a trained, experienced author?
Think of it this way: The internet has made research vastly more accessible. It has provided a shortcut TO the research; it isn’t writing it. And that’s how we can be ethical AI users, by using the tool, not the talent.
The Guardrails
A few weeks back I was putting together a piece of marketing content that I had written in one form or another at least 100 times. It’s so familiar to me, yet I couldn’t find the words to start. Maybe at this point in my career I have too many reference points.
So, I sat looking at the screen for about 10 minutes and then decided to plug my carefully engineered prompt into ChatGPT. Et voila! The results reminded me of what I wanted to say, not HOW I wanted to say it, necessarily, but it gave me the bones.
That’s Guardrail #1: Never let an AI-generated piece of content speak for you.
When you do that, you give it creative control. Let it gather and summarize, and then you should take the wheel to make it a great piece of content that reflects your expertise, insights and understanding of the assignment.
Guardrail #2: Make sure you can stand behind your work product.
AI can be famously wrong. In a process called “hallucinating”, AI tools can actually make things up when they don’t know the answers. So even if you get something out of ChatGPT that sounds great and on point, fact check it. Writers can find a lot of inspiration in AI tools but should never depend on their outputs. Credibility can be lost in an instant.
Guardrail #3: Be transparent.
I think the communications world and all the clients we support are quite aware that AI is in the mix and is likely used every day in some way, shape or form. That doesn’t mean they are excited about it. In fact, some of our clients are extremely wary about what it can mean for their organizations. At Vanguard, we are intentional about how we use AI. When it has been a companion to us in a work product, we say so. (e.g., “We used AI to help us brainstorm these campaign names.”)
For brainstorming purposes, AI is just a few steps away from plugging ideas into Google, so it feels like a natural and “safe” step for us. But we still always reveal it, and our clients appreciate it.
Our Value Continues
Organizations want to see words that reflect our experiences and knowledge, and that expertise goes deeper than what the internet can uncover. AI wasn’t there for the conversation I had with a colleague that made me pivot on a campaign direction. And it wasn’t part of the meeting I had with my client last week where they gave me feedback on an op-ed draft. Those are my experiences, which will lead to better drafts.
And that’s how I can sleep at night. AI is not omnipotent, but using it ethically is a muscle we are all going to have to strengthen. Right now, while large language models are still relatively new, it’s a good time to set up a healthy relationship with AI.
Here are some tips that helped me get comfortable with AI as a writing partner:
- Start small, giving whatever platform you like a few exploratory tasks—even personal ones that carry no professional risk. (I had ChatGPT develop an invitation to a summer solstice party I was throwing last summer, just to see where it landed and where it missed.)
- Watch webinars and videos on what AI can really do. I use it sparingly in client work but as noted earlier, it has some real potential to save time on marketing tasks, which feels like a better role for it.
- Be curious. To trust AI and use it well, you are going to have to lean into it. Research the pitfalls and know where you need to draw the line.
AI was not used in the development of this blog post.
LeAnne DeFrancesco is a vice president at Vanguard Communications in Washington, D.C., where she leads the firm’s Design and Editorial practice. She joined WWPR’s Pro Bono Committee in 2018, where she has helped several D.C.-based organizations enhance their PR and communications activities.
