Here at Message Lab, people constantly ask us how AI will change our jobs. The honest answer? We don’t know yet!
So far, AI is impressively good at writing simple content. But in the areas we write about — healthcare, scientific research, philanthropy — AI falls short of capturing the nuance, accuracy and strategic messaging we need. And it can’t really help with the trickiest part of our job: helping a doctor or researcher articulate exactly what their work is and why it’s important for their audience.
That’s why we don’t use AI for any actual content writing. Still, it’s a useful tool that helps us generate ideas and makes our overall writing process more efficient.
Here’s how our team is using AI in content marketing, some tips for how it might help you — and some pitfalls to avoid.
H2: How we use AI in content marketing
We think of AI as a personal assistant. We can use it to support simple tasks that take up time, so we can use our human brains for the tasks that need them most. We most frequently use AI in content marketing for:
Making complex things simple
AI can be a useful tool for generating metaphors and simple explanations across many types of content. We might use prompts like:
- What are three metaphors for how the kidneys work?
- Explain how the gallbladder works to a 5th grader.
We don’t use the explanation it provides word for word. But this can generate ideas for jumping-off points when explaining complex things.
We have also tested ChatGPT in summarizing highly scientific content in lay language. For example, we’ll paste the abstract of a complex scientific paper into ChatGPT and ask it to explain the paper in simple terms. Voila – it’s suddenly easy to understand a new finding about, say, autoantibody reversion, without the fits and starts of old-school Googling.
Still, while ChatGPT is great at capturing the overall gist of what we ask it to translate, it isn’t always accurate about the nitty-gritty details. That’s why we also talk to subject matter experts, so they can confirm when we’re understanding things correctly – and set us straight when we’re not.
Idea brainstorming
Ideally, your story or blog post ideas are coming from a clear strategy that’s aligned with your marketing goals. But it can still be easy to get stuck when brainstorming what to write about — especially if you have already written a lot about a topic. AI can help come up with ideas when you feed it prompts like:
- Give me some ideas for blog posts about asthma in teens.
- What are healthcare topics that people are talking about during the winter?
- What questions might parents have about type 2 diabetes?
- What are examples of how teens might practice self-care?
H3: Help with simple copy (that breaks your brain at the end of the day)
We don’t trust AI to write full articles or anything with concrete facts. But it can be a helpful jumping-off point for short snippets of straightforward copy, or to brainstorm ideas in places we get stuck like:
- Titles
- Email teasers
- Meta Descriptions
Non-generative AI tools for content marketing
Generative AI tools like ChatGPT get most of the attention, but two other AI tools that make our lives a lot easier are:
- AI transcription tools, which can provide cheap, fast and accurate interview transcripts.
- AI spelling and grammar checking tools, for a quick proofread.
Prompting in AI for content marketing
If you plan to use generative AI in the above ways, try these tips for writing your prompt:
- Be specific. For example, provide parameters about the length of response you’re looking for and the reading level of the content.
- Provide context. For example, rather than “explain the benefits of telehealth” tell it “I am a healthcare marketer working for a rural clinic that’s implementing telemedicine. Some of our elderly patients are reluctant to use this technology. Can you explain the benefits of telemedicine specifically for elderly patients in rural areas, and how it can improve their access to healthcare?”
- Ask for several options. For example, don’t just ask for a title for a blog post about nutrition for babies. Instead, ask for “10 titles for a blog post about nutrition for babies.” This will provide a lot of ideas to jumpstart your brainstorming.
- Ask follow-up questions or rephrase your question if you don’t get what you’re looking for the first time. Even simple prompts like “make it more conversational” or “add detail about X” can get you much closer to what you need.
Downfalls of AI is for in content marketing
We don’t use AI for any significant amount of writing for the following reasons:
- Inability to generate new information. We’re often asked to write about novel medical treatments and new research studies and ideas. ChatGPT isn’t very helpful with these cases because it pulls from existing information – it can’t (as far as we know) interview subject matter experts to gather the information needed to write articles, webpages or stories that don’t already exist somewhere online.
- Inaccuracy. Information written by AI may not be factual or correct.
- Quality of writing. The content often falls short of our standard in terms of voice, tone, clarity and quality. AI will likely get much better over time, but most AI-generated content just isn’t very engaging yet – at least not to real humans.
- Ethical and privacy concerns. It’s often unclear where the information that you get from AI comes from or if it’s proprietary.
- Negative impact on search. One recent study found that human-authored content improved traffic and rankings by three times more than AI-authored content and AI-authored content actually caused rankings to fall.
One nice thing about ChatGPT is that it’s usually honest about its limits. Here’s what it said when we asked: “What are some downfalls for using AI for content marketing?”
- Lack of Creativity and Human Touch: AI often lacks the emotional depth and cultural nuances necessary for truly original content.
- Over-Reliance on Automation: Dependence on AI tools can lead to generic content that loses unique brand voice and personalization.
- Limited Understanding of Context: AI may misinterpret context or audience needs, especially in sensitive fields like healthcare.
- Data Dependency: The effectiveness of AI is tied to the quality of its training data, and poor data can produce misleading content.
- SEO Saturation: Increased use of AI for SEO may flood search engines with similar, low-quality content, making it hard to stand out.
- Risk of Errors: AI can produce factual inaccuracies, posing risks in industries where accuracy is critical, like healthcare and finance.
We couldn’t have said it any better. Ok, maybe a little!
Learn more about Healthcare Content Marketing and Nonprofit Content Marketing with Message Lab Media
