Since you’re reading this, I assume you’re among the hundreds of millions of people hungry to learn about ChatGPT. And no, that number is not an exaggeration; I just searched “ChatGPT” on Google and got 230,000,000,000 results. It’s not apples to apples, but the interest is there.

We’ve seen this many times before: an intriguing tech innovation that changes the game. ChatGPT and its ilk (Jasper.ai, Phrasee, Smartwriter.ai) may be entering into a lasting category of functionality that marks this as a watershed moment.

Or there’s another scenario that should be familiar to anyone who’s ever hung up on a robocall or sent impersonal emails to spam: it could become a straight-up nuisance. Where things go from here isn’t about the technology; it’s about the humans engaging it. And here’s where things could go right — or wrong.

Related: How Can Marketers Use ChatGPT? Here Are the Top 11 Uses.

What is the ideal state for human/AI interaction?

AI (artificial intelligence) will almost never be great in a silo. Love or hate Peter Thiel’s politics, his position (which he shared in Zero to One) that technology and human expertise together will always be better than technology alone is one that I share.

As the tools come out, everyone starts using them. While they may briefly hit a plateau of utility and become the new lowest common denominator, humans who strive to make the tool’s output better will add value – especially if the tool, as ChatGPT appears to, refines its results with time and volume (much like Facebook’s ad algorithm, for example).

I hear a lot of chatter about ChatGPT “will change X as we know it.” And that might be true, but not necessarily in a bad way. Tech that proves truly useful and valuable — think of how instantaneously Facebook’s lookalike audiences changed the paid social game with their relevance — won’t take over as much as it will help us improve at our jobs.

Related: Chat GPT Urgent Warning: Is Your Information, Job or Business Safe?

What does a race to the bottom look like with ChatGPT?

Let’s go back to robocalls and emails for a second. You’ve gotten spam calls that start with a suspiciously long pause, then a robotic voice, right? And you’ve gotten spam emails that begin with “Hello,” not even your name, and then just jump right into a sales pitch.

Years ago, when those technologies were introduced, they were revolutionary. They were ways to reach actual people, at scale, in seconds. Today, those are recognizable as lazy tech applications with no human refinement. That’s a race to the bottom: people getting lazy and either overusing the tech, using it with no thought or enhancement, or both. And you can very clearly see this coming in the ChatGPT era: obvious query, un-nuanced result, cut/paste, post.

The good news is that soon enough, people will be able to recognize those results for what they are, and the power of the tech alone will be diluted. Early adopters might get wins now, but when the tech proliferates, the baseline usage will carry no value. Just as with email, it won’t mean that ChatGPT and its ilk are dead; they’ll only be valuable when used well, with humans at the controls and applying critical thinking.

Does ChatGPT post a threat to marketing professionals?

ChatGPT and other AI tools pose a potential short-term threat to folks who do things like content creation, video and image creation, and possibly keyword research since it can produce baselines of those initiatives quickly and for much less than you’d pay a professional.

That said, smart professionals might have an arbitrage opportunity of their own if they can use those tools to get to a certain level quickly, then add their expertise to produce uniquely high-quality results. (If you’re an expert copy editor, for instance, you can take a piece of content that Grammarly has already polished and bring the language itself to another, more sophisticated level that truly tugs on the reader’s emotions.)

Over the long term, ChatGPT and similar tools will become (or already are) part of the marketing toolkit. The best way to deal with it is to adopt it and adapt quickly — understand its strengths and how to use them and its weaknesses and how to improve on them. Put it in your own toolkit now, and learn how to evolve with it over time.

Powerful use cases for AI in marketing – right now

AI has been around in marketing for years. Famously, Facebook and Google advertising are governed by machine-learning algorithms, and Amazon has used AI to design personalized experiences for far longer. That said, personalization, in general, will continue to be an area where AI can provide value, particularly in determining the right content — the type of content and messaging — to put in front of users.

Whether it’s chatbot-produced text or AI-produced videos or images, AI’s power lies in its speed (create and use assets almost instantaneously, with no uploads or storage steps). Let’s say you’re Kellogg’s and advertising a kids’ and an adults’ cereal — you can personalize creative to show your adult audience ads for each: a table of happy kids eating the first cereal and a healthy-looking adult eating the second.

Essentially, if you get first-party data in the mix and understand how to use the demographic and interest-based information you have in tandem with AI, you can put effectively personalized creative in front of your users at scale without thinking about it. (Although once everyone’s doing that, the next layer of human creativity atop the AI-produced creative will be the differentiating factor.)

AI is also the backbone of predictive analytics and an important marketing sub-discipline, media mix modeling (which enables marketers to understand which channels should get their budget and what results they should expect). But while AI is the engine for the analysis, marketers need to provide the right fuel (data input) and steering (data interpretation and follow-up decisions and actions). In other words, marketers need to ask the right questions, assess the analysis to make sure it’s logical, and take action on the analysis by building and optimizing media campaigns.

Whether in marketing or other disciplines ripe for disruption by technology, few technologies have the power to change an industry’s trajectory. ChatGPT and other AI innovations are impressive on many levels, but their full potential will only be realized when smart people apply themselves to leveraging the technology to create a new productivity baseline.