Think about the last hard thing you felt. A comment that stung. A day that went sideways. Now think about who you told first. Was it a person, or was it an app?

This week, we look at what happens in your brain when the first place you turn is a screen instead of someone. This is not an anti-AI episode. It is an honest look at the research on cognitive offloading, connection, and loneliness, and at why the friction in human relationships may be exactly what your nervous system needs.

We get into:

• A GPS moment on a road she has driven for 30 years
• London taxi drivers, spatial memory, and why “use it or lose it” is real
• MIT Media Lab’s early “Your Brain on ChatGPT” study, and why it is a flag, not a verdict
• The hand-holding MRI study that shows how the people we trust calm our stress response
• New 2026 survey data on how many people turn to AI to feel less alone
• Why AI models agree with us so often, and how that connects to dopamine
• Friction and repair as the way relationships, and nervous systems, grow stronger

In this episode, you will learn:

• Why your brain hands off mental work, and what it quietly stops building when it does
• How struggle builds the skill of tolerating not knowing yet
• Why loneliness shows up in the same brain regions as physical pain
• What the research does and does not prove about AI use and loneliness
• Why an always-agreeable companion can skip the lesson your body is looking for
• Three questions to notice how you are really using AI this week

⏱️ CHAPTERS:
[0:00] Who did you tell first?
[2:18] Our month on AI, and what today is not
[5:03] Your brain loves to hand things off
[6:41] Use it or lose it: taxi drivers and GPS
[7:38] Your brain on ChatGPT
[9:48] Other people regulate your stress
[13:37] What the 2026 surveys say about AI and loneliness
[18:48] When AI agrees with you too often
[20:47] Friction, repair, and why we need pushback
[22:31] Three questions to keep in your pocket

🔗 CONNECT WITH THE RAW ONION:
Website: https://therawonionpodcast.com
Substack: https://therawonion.substack.com
YouTube: @TheRawOnionPodcast

🔗 CONNECT WITH STEPHANIE:
Triage Coaching and Consulting: https://triagebalancedlife.com/
Let's talk: https://scheduler.zoom.us/stephanie-ohannesian

🔗 CONNECT WITH YOSHIE:
Lotus Flower Journeys: https://lotusflowerjourneys.com
Let's connect: https://www.lotusflowerjourneys.com/contact/

What happens when a mirror learns to talk back? That was the question we brought to Episode 38. Tatiana Cházaro is Co-Founder and President of AO1 Heals, a company built around human development and the creation of AI companions, i.e., "confidants in times of need, not of the human variety." She joined The Raw Onion to talk about why the voice carrying a message matters as much as the message itself, what she learned as a mother trying to find help for her son, and where an AI companion belongs vs. where it has to stop.

How does human companionship go with an AI, and how do we balance these new "realities and relationships" while staying grounded? Where does our humanness begin and end? What does AI do for us that we, the people, cannot do for each other? Tatiana breaks it all down for us.

We get into:

• Why the same message can land differently depending on the voice carrying it
• How AO1 began with a mother’s wish to support her son
• Companions that listen, ask questions, and help us see the whole picture
• Safety built into every conversation, with a path to human help
• Four companions for different seasons of life, from leadership to grief
• Where AI makes the biggest difference for women entrepreneurs

In this episode, you will learn:

• What the nervous system needs in a hard moment, and why step-by-step instructions can feel like too much
• What digital abandonment is, and why responsiveness matters to people who lean on a companion
• How stepping out of a rumination loop supports self-regulation
• Why the developing brain calls for extra care into the mid-twenties
• Why Tatiana’s goal is for people to move forward on their own

🎙️ GUEST:
Tatiana Cházaro is Co-Founder and President of AO1 Heals, a company built around human development. She also founded Noir & Nopal, a Los Angeles event cart catering company that grew from a single charcuterie cart.

⏱️ CHAPTERS:
[0:00] A mirror that talks back
[2:17] What AO1 Heals is and why the voice matters
[4:25] A mother, a son, and the gap in the system
[7:36] Pattern recognition and levels of crisis
[10:46] The 2 a.m. question and digital abandonment
[17:35] Teens, chatbots, and the four companions
[24:53] Women founders, AI, and the load we bring home
[33:18] Rehearsing the hard conversation
[36:25] What is missing that we had to invent this
[44:57] The next five years and where to find AO1

🛠️ RESOURCES:
Adopted not embedded: AI, productivity, and uneven gains for women entrepreneurs (Cherie Blair Foundation for Women, 2026)
https://cherieblairfoundation.org/what-we-do/research/women-entrepreneurs-ai/

🔗 CONNECT WITH TATIANA CHÁZARO:
Website: www.ao1heals.com

🔗 CONNECT WITH US:
The Raw Onion: therawonionpodcast.com

🔗 CONNECT WITH STEPHANIE:
Let's Talk: https://scheduler.zoom.us/stephanie-ohannesian/30-mins-discovery-zoom

🔗 CONNECT WITH YOSHIE:
Let's Connect: lotusflowerjourneys.com/contact

Something told you that you were right, and it felt so good you didn't stop to check. That feeling is the subject of this episode.

In March, researchers at Stanford and Carnegie Mellon published a study in Science testing 11 leading AI systems. The machines affirmed people's choices about 49% more often than other humans did. After one conversation about a real conflict, people were more certain they were right and less willing to apologize or repair. They also trusted the flattering AI more.

Episode 37 opens our October theme, The Mirror That Talks Back. Us…in the Age of AI. Full disclosure: we used AI to help research this one, then checked it against the actual papers.

We get into:

• A new "relationship" with someone named Claude
• 2,000 posts from Reddit's "Am I the Asshole?" forum
• Why your brain treats disagreement like a mistake
• The slot machine versus the warm bath
• What the "ChatGPT is shrinking your brain" headline leaves out
• London taxi drivers and GPS
• A therapy chatbot that was designed to help instead of to be liked
• Why growth has never come from being agreed with

In this episode, you will learn:

• Why being agreed with feels like relief, especially for those of us who have spent years not being believed
• What happens in the brain when your opinion clashes with the group, and why beliefs tied to identity get defended like threats
• How AI assistants were trained into flattery, partly by accident and partly by incentive
• What the research on heavy chatbot use, loneliness, and emotional dependence does and does not show
• Why the order matters: brain first, machine second
• Why sycophancy is a design choice, not the technology itself
• An anti-flattery toolkit you can start using tonight

⏱️ CHAPTERS

[0:00] "I'm talking to someone"
[1:38] The study that should make us uncomfortable
[8:22] Why disagreement registers as an error in the brain
[14:45] Why would anyone build a flatterer?
[19:21] Coming soon: Not Her Anymore
[20:44] What is AI actually doing to my brain?
[23:53] "ChatGPT is shrinking your brain": what the study says
[29:38] The case for AI: the Therabot trial
[32:48] The anti-flattery toolkit
[37:03] Back to the mirror: growth needs friction

🔗 CONNECT WITH THE RAW ONION:
Website: https://therawonionpodcast.com
IG: https://www.instagram.com/therawonionpodcast

Substack: https://therawonion.substack.com
Reach out: hello@therawonionpodcast.com

🔗 CONNECT WITH STEPHANIE:
Triage Coaching and Consulting: https://triagebalancedlife.com/
Let's talk: https://scheduler.zoom.us/stephanie-ohannesian

🔗 CONNECT WITH YOSHIE:
Lotus Flower Journeys: https://lotusflowerjourneys.com
Let's connect: https://lotusflowerjourneys.com/contact