TWO AI EXPERIMENTS: A CREATIVITY BOOST OR A SHORTCUT TO BLAND?

Two AI experiments caught my eye this week in the NYT, each with wildly different outcomes.

What’s the common denominator? Human input.

🎭 FIRST UP, ROBERT DOWNEY JR. ON BROADWAY

The award-winning actor stars as a literary giant in a play where the entitled protagonist decides to let AI write his next book.

Here's the kicker: the real-life playwright used ChatGPT to refine a speech he wrote, and when he read it back to the cast, "jaws dropped." Early critics conclude that the play achieves its goal of creating a “bewildering and unsettling experience.”

(Only two of the chatbot's lines survived in the final play.)

📝 ON THE FLIP SIDE: A JOURNALIST LETS AI RUN HER LIFE FOR A WEEK

Yip—all her daily decisions, from what to eat, wear, and do with her kids, to what haircut style to choose.

But all the context she provided was this: “I was a journalist conducting an experiment and that I had a family.” Unsurprisingly, the AI churned out choices so predictable they felt, well, flat—“erasing individuality in favor of a bland statistical average.”

🔍 MY AI EXPERIENCE: YOU GET WHAT YOU PUT IN

What I’m seeing here mirrors my experience: AI is just a tool, and how “good” it gets is all about how you use it.

When I take the time to feed it context, share my thoughts, and go back and forth a few times, that’s where I find AI most interesting—fueling and expediting a thoughtful creative process that allows me to explore, ruminate, and deepen my ideas. And I always (okay, almost always!) take over for the final draft, making sure the result sounds like me.

But toss it a vague request without context or my own ideas? That’s when I get “generic” and “blah.” Sometimes, “generic” is all I need—a quick response to that pesky salesperson who, nonetheless, I don’t want to alienate!

🎙️ ANOTHER AI EXPERIMENT WORTH THE LISTEN—A JOURNALIST WHO "CLONES" HIMSELF

In veteran journalist Evan Ratliff’s Shell Game podcast, he’s testing how far he can replace himself with an AI voice clone, including activities that many assume AI can't do yet, like meetings and interviews. Turns out his AI clone is surprisingly effective (if not always convincing!), challenging our assumptions about what AI can and can’t do.

Whether you see AI as a creative muse or just a quick fix, it’s all about how you use it.

Who’s experimenting with AI in their work? What’s surprising you—or leaving you underwhelmed?