If you’re expecting too much from generative AI tools and tech, you’re not alone.
Each week of AI hype and announcements brings a fresh wave of anticipation and … fear.
For instance, OpenAI’s newest model, ChatGPT-4.5, promises more personality or ‘vibes’ than other iterations. But what does that mean?
Then, there are all the reasoning models, like DeepSeek, OpenAI’s o3, Gemini 2.0 Flash Thinking and Anthropic’s Claude 3.7 Sonnet. They’re designed to think through problems slowly and to show their work, so they’re better at coding, math and other logic-based problems, but may not be as creative as other LLMs.
Yesterday’s Axios A+ newsletter described the situation as the classic high school battle between the ‘freaks and geeks’. Between language and math. Creativity and logic. A tug of war between two distinct personality types.
They wondered how long it might be before generative AI tools could blend both.
But until then, which ones should you choose?
How Do You Measure the Value of AI Tools?
That’s a question on many marketers’ minds.
According to a Gartner survey from last fall, 41% of CMOs experienced resistance from their teams when it came to adopting generative AI. This was due, in part, to employee concerns about being replaced by machines.
And it occurred to me that when you look at the promise of what AI can do versus the fear of becoming obsolete (FOBO), there’s a duality that underpins every aspect of artificial intelligence.
Every generative AI tool has two sides: positive and negative features and consequences. And there seems to be a constant battle between them to gain the upper hand.
Which means that whenever you use an AI tool, you experience this conflict first-hand.
And the push and pull that’s throwing us off-balance, could be responsible for your uncertainty and stress.
Understanding the Tradeoffs
So how do you find an equilibrium?
Let’s begin by looking at the plus and minus sides of AI use. Here’s a chart to illustrate what I mean.
On the left, you have positive aspects from the work marketers and comms pros produce. On the right, the negatives.
Because Generative AI is grounded in language and communications, it has a similar range.
Its plusses include an ability to quickly come up with an endless stream of ideas (including many stinkers), edit text and images, produce visual concepts, have conversations and provide alternate points of view. It’s also a good platform for A/B testing.
But the negative side, if left unchecked, could lead to catastrophes like blatant errors, mass persuasion, disinformation at a scale we’ve never before seen, biases, and an inability to explain its thinking or answer why.
And while many of today’s AI outputs are less than stellar, its overall performance is improving in leaps and bounds, in part because we’re training it with our data every time we use it.
How Do You Strike a Balance?
I wish I had the answer. But that comes down to an individual choice only you can make.
One thing you can do is consider the +/- chart like a dashboard, where you adjust the settings to each of the parameters when you use AI, or as a rationale for the times you don’t.
Let’s use this newsletter as an example. I like writing! So while AI may be faster, I still have a big enough ego to believe I’m a better writer than a machine. So I’ll strive for originality over cliche and that means most of the heavy lifting falls to me.
Creativity is another place I’m not willing to make tradeoffs so I would take on most of the responsibility and not rely on AI. I would also dial up safety.
For each of those I’d set the human dial to say, 95% (or higher).
But as we move down the list, my settings become more nuanced. For instance, I’m not as concerned with my own privacy on social and digital media and would dial up the public availability of my content. Maybe 70% public versus 30% private.
For authentic vs. synthetic, I might set it closer to the middle as I often use text-to-image tools for visuals. But I also disclose.
And depending on whether the topic I was writing about was more speculative or based on facts, I might increase either human feeling or data.
These are just a few examples. But they give you a place to begin.
The duality of AI isn’t about finding a simple answer.
Nor is it sticking to the middle of the range.
Each of you has to decide which tasks you want to keep for yourself (because you enjoy doing them) and which you’re willing to share with a machine.
And of course, that will shift over time.
On the human/AI collaboration spectrum, where do you draw the line?
Connect with FMI
Want to stay ahead of the marketing wave and prepare yourself for the future? Connect with us on our Website, LinkedIn or Instagram.
If you enjoyed this newsletter, please share it and subscribe to receive it directly in your inbox.
This post was written by FMI Associate Director, Martin Waxman. A version of this post was originally published in Digital Marketing Trends.