Overview of Where Should We Begin? — “Is Using ChatGPT Hurting My Relationship?”
In this episode, a caller asks Esther Perel whether using ChatGPT to navigate conflict in her emotionally intense, long-distance relationship is a helpful self-regulation tool or an unethical breach of trust. The conversation explores how AI became a real-time conflict coach, how it helped her break reactive loops, and whether secret use of the tool undermines intimacy, transparency, or the integrity of the relationship itself.
What the Episode Is About
The caller is in a relationship with a partner in recovery who can become highly reactive, especially over text. After years in a stable, low-conflict marriage, she felt out of her depth and turned to ChatGPT for support.
She used AI in two ways:
- As a reflection tool to understand the relationship dynamic
- In real time during conflict to help regulate her responses and break escalation loops
Over time, she felt the tool helped her become:
- Less reactive
- More grounded
- Better at setting boundaries
- More able to tolerate difference without escalating conflict
But she also wrestled with the fact that:
- Her partner does not know she used AI this way
- He is explicitly anti-AI
- She sometimes copied and pasted their conversations into ChatGPT
- She wonders whether this is private self-help or a form of secrecy that compromises intimacy
Key Relationship Dynamics Discussed
Conflict, reactivity, and looping
The relationship is marked by:
- Very high highs and very low lows
- Fast-escalating arguments, mostly over text
- A pattern of blame, defensiveness, and circular conflict
The caller initially tried to:
- Over-explain herself
- Prove she was not doing what he feared
- Legitimize her actions and defend her intentions
Esther points out that these strategies kept her trapped in the loop.
Trust, suspicion, and perception
A major conflict example involved a male friend staying at her house while her partner was away. He accused her of cheating, and she instinctively tried to prove her innocence.
Esther reframes the issue as not just about the specific event, but about:
- Trust
- Interpretation
- Whether her partner sees her perception as credible
This becomes a larger issue of epistemic trust: does he believe her judgment is sound, or does he consistently place his own read above hers?
Differing worldviews and power dynamics
The couple also differs strongly in:
- Politics
- Trust in institutions
- Belief in conspiracy theories
- Expectations around information sources
The caller says he sees her as naive and overly trusting, while she sees him as suspicious and more “in the know.” Esther notes that this creates a subtle but important hierarchy: one person becomes the “more accurate” one.
Esther Perel’s Main Insights
1. AI was coaching the relationship, not just the individual
Esther makes a key distinction: the caller was not simply using AI for personal improvement. She was using it to change the interactional pattern between two people.
By changing her responses, she changed the loop, which changed how he responded.
2. The tool was teaching differentiation
Esther argues the AI was helping the caller tolerate difference without collapsing into defensiveness or fusion.
In other words, it helped her:
- Hear him without agreeing with him
- Hold her own boundaries
- Resist constant self-legitimation
- Stay present without becoming reactive
3. The real question is not “AI or no AI,” but “what kind of support, and for what purpose?”
Esther compares AI use to:
- Therapy
- Melatonin
- Diet changes
- Any personal practice that one partner may not like
Her point: the bigger issue is the relationship’s tolerance for difference and the amount of reactivity to something one partner disapproves of.
4. Privacy and secrecy are not the same thing
The caller came into the conversation believing intimacy meant full disclosure. Esther complicates that idea by distinguishing:
- Privacy: what belongs to one person
- Secrecy: hiding something in a way that undermines trust
She also notes that in a long-distance relationship, the boundaries of disclosure may naturally differ from an in-person, cohabiting relationship.
5. In recovery dynamics, mistrust often runs deep
Esther points out that people in recovery may be especially sensitive to concealment, because addiction often involves hiding and mistrust. That can heighten suspicion in a relationship, even when the actual issue is not about deception.
Notable Takeaways
- Changing yourself can change the relationship loop.
- AI can function as a tool for self-regulation, boundary-setting, and perspective-taking.
- The ethical tension is not just about the tool itself, but about secrecy, trust, and the couple’s agreement around privacy.
- The caller’s real challenge is not only conflict management, but also whether the relationship can sustain difference without assigning one partner a lower level of discernment.
- Intense chemistry can keep a relationship alive, but it cannot substitute for respect, trust, and mutual recognition.
Questions Esther Leaves the Caller With
Esther does not give a simple yes/no answer. Instead, she asks the caller to reflect on:
- What counts as privacy in this relationship?
- What kinds of things should be shared, and what can remain personal?
- Does the secrecy around AI matter because of the tool itself, or because it mirrors broader trust issues?
- Does the relationship support or erode her ability to trust her own perception?
- Is she being asked to smooth over differences at the cost of being diminished?
Bottom Line
This episode is less about whether ChatGPT is “good” or “bad” in a relationship and more about how a new tool can become part of an existing relational system. Esther frames AI as a powerful, practical support for breaking conflict loops — but also raises the deeper question of whether using it secretly reflects a broader intimacy issue around transparency, difference, and trust.