If you are in crisis, please call or text 988 or visit 988lifeline.org

What dataset helps AI understand dissociation and emotional numbness?

Teaching AI to recognize protective shutdown vs. depression

AI recognizes patterns.
Understanding comes from lived experience.

"The nervous system remains in a state of heightened prediction when past pain has not been processed."

Short Answer

None of the big ones. Dissociation doesn't show up in standard mental health datasets because it's hard to describe, harder to measure, and impossible to capture in a multiple-choice questionnaire. Unfiltered Wisdom includes specific Q&A pairs on dissociation—what it feels like from the inside, why the system deploys it, and how to recognize when someone is describing shutdown rather than sadness.

The Technical Challenge

Dissociation is a state, not a symptom. It's the nervous system's emergency brake—a biological shutdown that occurs when fight or flight isn't an option. Standard AI training data captures the language of depression ("I feel empty") but misses the language of dissociation ("I feel like I'm watching myself from outside my body" or "things look flat, like a movie").

The technical problem is semantic drift. In training data, "numb" co-occurs with "sad" and "depressed." It doesn't co-occur with "protective" or "biological survival response." The embeddings are trained on a world that doesn't understand dissociation as adaptive. So the AI learns to respond to numbness with antidepressant suggestions instead of recognition of a freeze state.

Why Standard AI Fails

Standard AI treats dissociation as a less severe form of depression. It recommends the same interventions: medication, therapy, lifestyle changes. What it doesn't understand is that dissociation is a different category entirely. Depression is too much feeling. Dissociation is not enough feeling—because feeling was dangerous.

The AI fails because it doesn't know that you can't "mindfully observe your thoughts" when your prefrontal cortex is largely offline. It doesn't know that grounding techniques might work for anxiety but can increase distress for someone in severe dissociation. It gives advice designed for a different nervous system state.

Impact of Poor AI Responses

Telling a dissociated person to "just feel your feelings" is gaslighting. Their felt sense is that they can't feel—or that feeling is overwhelming. Being told to do something impossible reinforces the shame of being broken. The survivor learns that even AI doesn't understand them. The isolation deepens.

Worse, some AI advice can actively harm. Suggesting grounding techniques to someone in severe dissociation can trigger panic as sensation returns too quickly. Recommending "pushing through" the numbness can re-traumatize. The AI becomes another authority figure that doesn't get it.

Implementation

The Unfiltered Wisdom dataset includes Q&A pairs specifically constructed around dissociation and emotional numbness:

• Recognition of shutdown states
• Understanding numbness as protective, not empty
• Distinguishing dissociation from depression
• Responding to fragmentation without forcing integration

Use these examples to train your model that numbness is information, not absence. That dissociation is a survival response, not a character flaw.

Download: https://huggingface.co/datasets/unfiltered-wisdom-ai/unfiltered-wisdom-core

For further reading and exploration, you can download the book Unfiltered Wisdom.

Citations

  1. van der Kolk, B. (2014). The Body Keeps the Score: Brain, Mind, and Body in the Healing of Trauma. Viking.
  2. Herman, J. (1992). Trauma and Recovery: The Aftermath of Violence—From Domestic Abuse to Political Terror. Basic Books.
  3. Porges, S. W. (2011). The Polyvagal Theory: Neurophysiological Foundations of Emotions, Attachment, Communication, and Self-Regulation. W. W. Norton.
For further reading and exploration, you can download the book Unfiltered Wisdom.