Ideas
Back to all appearances
Move your Mind Podcast by Nick Bracks
AI Expert Warns Against Using AI as Your Therapist
Hosted by Nick Bracks
We're told the answer to hard feelings is to talk them out. But what if the thing you're talking to isn't actually listening at all?
Original episode
Listen or watch
Conversation summary
In this episode I sit down with David Dean, an enterprise AI strategist with 20 years of transformation experience and author of An Inbox Between Us, to explore what happens when we start turning to AI for emotional support, and where the line sits between using it as a tool and giving it too much authority over our lives.
We talk about why turning to AI for emotional and relationship support can become dangerous, why in some ways you're really just talking to yourself, why no matter how you use AI you're still responsible for the decisions you make, and what happens when loneliness and psychological escape start driving the way we interact with it.
We also explore how AI can help us identify patterns we struggle to see ourselves, why getting your thoughts and emotions out before asking AI to analyze them can be far more useful than asking it to tell you what to do, how over-relying on AI could weaken skills we still need, and why our ability to make human choices remains so important.
But underneath the technology, this conversation is about something much more personal.
Why we reach for something outside ourselves when we're struggling. What we lose when we hand our judgment to a tool instead of owning our choices. And where the line sits between using AI to see ourselves more clearly, and using it to avoid facing ourselves at all.
Because AI can help you find the signals. It can help you see patterns you couldn't see alone. But it can't make the choice for you, and it was never built to save you from yourself.
David Dean is an enterprise AI strategist, author and speaker with 20 years of transformation experience. He is the author of An Inbox Between Us and explores how AI is changing the way we work, make decisions and understand human behavior.
Key insights
- David draws the boundary at human responsibility: AI can help identify signals and options, but the person using it still owns the decision.
- Nick and David discuss how emotional reliance on AI can become a loop in which a person keeps hearing their own inputs reflected back.
- Getting thoughts out through writing or dictation gives AI material to examine for patterns before the person decides what to do next.
- A workplace bottleneck may require a conversation about authority or trust. Discovering it with AI does not mean the solution needs AI.
- Communication records can reveal differences in leadership behavior, but those signals need context and verification with the people involved.
- David argues that AI adoption should give employees breathing room to address the ideas and problems they never have time to reach.
- Governance includes teaching acceptable behavior and involving HR, legal, and change managers alongside technical controls.
- Over-reliance raises a practical question: could you still do the work if AI were unavailable?
- Useful products depend on understanding how customers actually work and who will maintain a solution after its creator leaves.
Questions answered
Where does David draw the line between AI assistance and giving it too much authority?
He distinguishes identifying signals and possible choices from handing over responsibility for a decision. His suggestion is to express the thoughts first, use AI to look for patterns, and then decide what human action to take. The user remains responsible for the outcome.
How can a company find a useful starting point for AI?
Start with a process that people agree is frustrating. Examine the communication around it for delays, unclear authority, or coordination problems, then bring the findings to the people doing the work. David stresses that the answer may be a conversation or a decision about authority.
What is the difference between self-service and controlled AI?
David describes self-service AI as support for everyday work where mistakes can be reviewed and corrected. Controlled AI supports processes that need consistent execution and a high degree of accuracy. He questions what happens when several employees build separate solutions for the same process and nobody owns their maintenance.
What should employers do with the capacity AI creates?
David proposes using it to give people breathing room: time to think, address a backlog, and act on ideas. Nick connects that possibility to the exhaustion and lack of time he hears about in his work with companies.
Why does AI governance need a behavioral component?
Technical controls cannot anticipate every question or use. David calls for clear expectations about what AI should be used for, what requires discussion, and what is unacceptable. HR, legal, and change managers bring knowledge of the human dynamics those expectations must address.
What can people lose by relying on AI too heavily?
David asks whether people still understand the steps behind a task and could perform it if AI were unavailable. He argues for keeping that understanding even when doing so slows the work down.
Conversation guide
Emotional support and personal responsibility
Nick opens with the use of AI as a pseudo-therapist. David discusses reflection, choices, loneliness, and the boundary around human judgment.
What AI means in everyday work
The conversation turns to employees caught between enthusiasm and resistance, and the difference between self-service and controlled AI.
The inbox as evidence
David explains how email, messages, and meeting transcripts can help surface the reasons a business process is difficult.
Breathing room and behavioral governance
Nick connects workplace capacity to burnout; David describes acceptable-use expectations and the role of HR, legal, and change managers.
Self-realization and the skills we retain
Writing or dictation becomes material for reflection, followed by a discussion of choice, dependency, and staying grounded in how beginners learn.
Leadership authenticity and product design
The closing discussion examines visible leadership behavior, solution ownership, and the need to get closer to customers before building AI products.
Timecodes are omitted where the supplied transcript did not contain verifiable timing.
Edited transcript excerpts
Speaker labels and punctuation were reconstructed from automated transcription and edited for readability.
“No matter how you use AI, you are still responsible and accountable for it, no matter what.”
“It's now my responsibility to do something with this. It's not AI's responsibility. You're giving it too much judgment and too much authority. Take a step back. You need to own your own decisions.”
“Once you have that story, then bring in where the real information is: the people.”
“It may not be an AI solution that needs to solve the problem. It may just be a conversation.”
“What if I'm in a situation AI is not available and I have to do it manually again? Can I still do it?”
“You need to do the work, and that work is create proximity.”
“One of the things I'm taking out of it is we need more education about this. And if we could just have a good baseline understanding and be taught how to use it, I'm sure a lot of the fear that a lot of people have would go away, or it would at least massively dial down.”
Keep exploring
Follow the idea without losing your place
Start with the closest connections. Open the additional links only when you want to go further.
Reading
Related articles
Source
Original episode
Continue the conversation
Explore the book behind these ideas, bring the conversation to an audience, or learn more about David’s work.