Almost every marketing practitioner I know is racing to figure out the best way to leverage ChatGPT for a competitive edge. From SEO to paid social to influencer marketing, ChatGPT is all the rage, as this Google Trends data shows:
That said, the rush to find powerful use cases must be balanced with an understanding of ChatGPT’s limitations and usage requirements. Before you use ChatGPT for something like keyword research or structured data construction, let’s dig into what I view as the five most important caveats to keep in mind.
1. ChatGPT is programmed to avoid certain kinds of content
Specifically, ChatGPT is programmed to not generate text on the topics of graphic violence, explicit sex, and potentially harmful content such as instructions on how to build an explosive device. On the whole, this is unquestionably a good thing that serves as a bit of a safeguard for brand integrity. But marketers should be aware that a) it may not cover every potentially harmful scenario; b) it may include some edge cases that affect industries like CBD, gaming, and pharma.
2. ChatGPT isn’t up on current events
My understanding is that, as of March 2023, ChatGPT is only working with information as recent as 2021. If your content needs to be up-to-date and fresh, then ChatGPT in its current form may not be useful. It’s likely that ChatGPT will get better at consuming current data in the near future, so keep an eye on this.
3. ChatGPT has built-in biases
ChatGPT is trained to be helpful, truthful, and harmless. Those aren’t just ideals; they are intentional biases that are built into the machine. It seems like the programming to be harmless makes the output avoid negativity. That’s a good thing, but it also subtly changes the article from one that might ideally be neutral.
The other caveat here is that ChatGPT results carry the implicit (or explicit) biases of the people who built the tool. Make sure to cast a critical eye on any results the tool gives you and look for potential bias to address as needed.
4. ChatGPT requires highly detailed instructions
ChatGPT results are only as good as the inputs that trigger them. The more detailed and nuanced the instructions, the higher-quality and sophisticated the output. For example, asking for “the 10 most popular French recipes” is fine, but “the 10 most popular classical French dishes made vegan” will get you much more specific results.
In other words, don’t get lazy. The fewer instructions you give in the request for content, the more likely that the output will look like everyone else’s, which adds literally no value.
5. ChatGPT content will be recognizable for what it is
OpenAI researchers have developed cryptographic watermarking that will aid in detection of content created through an OpenAI product like ChatGPT. Again, this is good news for people who want to use ChatGPT for good purposes and not just lazy, spammy results.
I foresee that ethical AI practices such as watermarking will evolve to be an industry standard in the way that Robots.txt became a standard for ethical crawling.
Keep these in mind as you evaluate how big a factor ChatGPT should be in your execution of marketing campaigns. If you have any questions about how we can use ChatGPT to help optimize your organic search, drop us a line.