Better Ideas, Faster: Using AI Creatively While Staying Ethically Grounded
Practical ways to boost productivity with AI while keeping authorship and ethics intact
Hi friends,
Today I want to talk about how to use AI as a creative, generative tool and the power of working with it iteratively to produce something you’re proud to put your name on. As always, I’m also thinking about the ethical dimensions of this work — especially from a legal ethics standpoint — so I’ll share some of the lines I’ve drawn for myself in using AI.
Over the past few months, I’ve come to a central thesis: AI is a powerful tool, but it is not a replacement for what we do as lawyers or as individuals. To me, that means two things:
Maintain initial authorship in the creative process so what we produce remains “our work,” not AI’s. This matters under the Rules of Professional Conduct and to ensure we use AI to make ourselves better — not lazier.
Remain the final editor. Even when AI helps with drafting, we must ensure that what we publish reflects our own knowledge, judgment, and intent.
Starting with the Right Inputs
This morning, I had a perfect example of using AI in this way. I was tasked by a company I do contract teaching for in my spare time to create an icebreaker for its annual retreat.
If I had simply typed into ChatGPT, “Create a fun icebreaker for two dozen people,” it would have produced something passable but generic. That approach would also raise two problems:
I’d be surrendering too much responsibility to AI, raising the ethical concerns I’m always thinking about.
I’d be missing an opportunity to get the best from both myself and the AI.
Instead, I gave ChatGPT context about the group (without sharing any confidential or private details) and the objectives I wanted to achieve:
Align with the retreat theme of adapting to change in our field
Show participants who’ve never used AI how it works
Keep it introvert-friendly (because I dislike small breakout groups myself)
Building Iteratively
The first draft ChatGPT gave me wasn’t “done” — and that’s the point. We started with a seed of an idea, then I told it what I liked, what I didn’t, and what I wanted to add. One key addition was finding a way to show participants what iterative AI collaboration looks like in action.
That’s how we developed “From Fragments to a Feature Story.”
In short: participants will use Slido.com to anonymously submit three quick facts about themselves. We’ll feed those into ChatGPT to generate a playful “Meet the Team” introduction, then improve it live based on audience feedback. We’ll end with a fun “2030 version” imagining our future success. It’s short, high-energy, and demonstrates adaptability, creativity, and the power of refining ideas collaboratively.
Along the way, ChatGPT even suggested a feature of Slido I hadn’t used before — so I asked it to explain exactly how to set it up. This is another quiet superpower of AI: it can give you clear, instant guidance on tools or processes you’re unfamiliar with.
Deciding When AI Should Be the First Author
Once we finalized the icebreaker, I needed to send an email to the organization’s leadership explaining what we’d developed and why. I gave ChatGPT the parameters, and it wrote a clean, professional draft.
Here’s the difference: for a routine administrative task like that, I’m comfortable letting AI be the “first author.” But for something like this Substack post — where I’m presenting myself as a creative writer — I always write first and use AI only as an editor.
That distinction is critical. Many lawyers have gotten into trouble by letting AI draft substantive work without their own authorship or review. For core functions, it’s a mistake to delegate to AI entirely. For auxiliary tasks, it can be a game-changing efficiency tool.
Where I’ve Applied This Process Before
I’ve used this collaborative, iterative approach with AI for everything from brainstorming Halloween party themes to helping draft explanatory memos that clarify work I had already created.
I once read that AI increases productivity by an average of 17%, and that feels right. If you use it well, you can produce better work, faster — without compromising ethics or authorship. AI can’t and shouldn’t replace us in our most essential work. But when used thoughtfully, it can help us deliver an exceptional final product.
If you found this post useful and want more practical, real-world examples of using AI ethically and creatively in law, teaching, and beyond, subscribe now so you don’t miss future posts.