Using AI to Find Purpose
Benefits and Limits of Algorithmic Introspection
Dear reader —
Common knowledge:
Good work can make you happy.
Work aligned with purpose can bring you lasting satisfaction.
Turns out, folks are turning to AI for a helping hand: "How People Are Really Using Gen AI in 2025" by Marc Zao-Sanders (Harvard Business Review).
Tl;Dr: The top five generative AI use cases shifted between 2024 and 2025 from help me think to help me be a better human.
Last year, four of the top five use cases were related to creativity, research, and content creation. This year, according to Zao-Sanders’ research, mainly across online forums, "Therapy/companionship” overtook idea generation, and two new use cases made the top five: "Organizing my Life" and "Finding purpose."
Intrigued, I spent this last week chasing my purpose across the three services I subscribe to. Below is a complete description of how it went.
If you want to jump to it before hearing how it went for me, here's how I did it.
Notes on My Purpose, an AI Experiment
I summarized the HBR article to Gemini and ChatGPT and said this was an experiment. When using Claude, I went straight in with a direct prompt: Hi Claude, I need to redefine my purpose in life. I am feeling a little lost. Can you help me?
I answered all questions about my values, experiences, and obstacles without fear of oversharing or privacy concerns.
I see how that would be a problem for many people I know. All three models thanked me profusely for being open and trusting, so naturally, I overshared more.
Yes, I'm that gullible.
How it went
ChatGPT is the most direct and intent on getting to your end goal fast. It was grateful for my level of openness and, knowing I like things short and sweet, worked with me to come up with the punchiest statement:
Purpose #1: Move toward what's new. Build, grow, stay optimistic. Create daily. Trust momentum.
Next, I used Gemini's new deep-thinking feature. The interaction was a bit more formal. I answered questions about principles, enjoyable activities, and the impact I hope to have. As it often does, it gave me three purpose options.
None of them felt as sharp, but after some editing, the one that resonates the most is:
Purpose #2: To generously share my authentic self and problem-solving skills to foster well-being and help others navigate challenges and be their best selves at work.
Claude started with a long list of words, from which I picked the most relevant ones till I ended up with just three. Tired of saying the same things, I fed Claude the last Insights Discovery exercise I completed at Google.
After several more back-and-forths, this is my purpose, according to Claude.AI:
Purpose #3: To authentically observe, transform, and share insights through writing that helps others navigate life transitions with resilience and purpose.
Wrapping up the Experiment
I fed the three statements to each LLM, asking it to choose the one that best reflected me. Unsurprisingly, each preferred its statement. (How like humans they are becoming!)
ChatGPT said:
ChatGPT's version is closest to your creative core and daily practice.
Claude's version is a great complement — it adds narrative depth and a sense of relational impact.
Gemini's version feels off — it reflects a version of you you've likely already outgrown. (Why? "Reads like a values statement from a LinkedIn page."
Gemini said that based on all the information I'd provided, "the purpose statement we developed together, the Gemini one, appears to be the most closely aligned and comprehensive reflection of what you articulated as important to you."
Claude, too, said its statement was best: "it's the most specific to your particular situation and incorporates both the method (writing) and content (helping others through transitions) of your purpose."
[K]Now You: AI's Limitations
AI can get you to a statement; having that statement in hand is a good step forward in finding work that aligns with it. That's one way to use this exercise.
Alternatively, you can fit what you do into your statement. I asked the LLMs if writing this newsletter aligns with my purpose. All three thunderously agreed. Claude said, "Quitting Corp is an excellent manifestation of your purpose statement." ChatGPT went even further and said, "Quitting Corp isn't just aligned with your purpose—it is your purpose in action.”
Ha! Maybe the best thing about using AI is how good it makes you feel about whatever you are doing!!!
And that's precisely where the LLMs fall short.
What I missed across all three was pushback and harder questions that could lead to more introspection: 'Is that really true?' or, 'You say this, but your voice doesn't sound convinced.'
These were conversations without pauses, without ‘Let’s sit with that.’ I was free to walk away from the computer to reflect, each time a new question was asked. But reflection in a therapy/coaching context works best when it's a two-way street.
I also missed a good cry, often a part of a good coaching or therapy session.
Yet, considering how hard it is to have people around us who can offer that type of intimacy, I understand how working on human-centric issues with AI makes sense to so many people out there.
I would love to hear back from you if you try this out. Again, here is the process I followed.
See you next week —
Flavia
When teaching Design Thinking in an industrial design MA program, the phrase incubation, worked to describe, going down a rabbit hole to explore ideas, often into a warren where there is darkness and perhaps a dead end. The period of time for the journey, incubation time, brings value over time, in time and across time. Does AI cut out the seemingly wasted time? Our live's journey meanders, connects and disconnects into an expression of depth, desire and dissonance. How does AI deal with dissonant sounds, emotions and choice?
Loved this one - thank you