Early-stage research • Built with care

AmygdalaHijack.AI
Building AI to detect and support recovery from human autonomy loss helping people regain control

We’re exploring how AI + neuroscience can identify early, pre-verbal signals of an Amygdala Hijack, the moment when autonomic responses override choice and reasoning. This is an early-stage research initiative. We do not provide medical advice, diagnosis, or treatment.

Note: AmygdalaHijack.AI is under development. Content here is for information and research collaboration only and is not medical advice, diagnosis, or treatment.

Autonomy Early Warning Neuroscience Ocular Signals Autonomic Regulation Visual Biomarkers Closed-Loop AI Real-Time Detection Non-Invasive Smooth Pursuit Parasympathetic Activation Nervous System Balance Signal-Driven Self-Regulation Human Autonomy Autonomy Early Warning Neuroscience Ocular Signals Autonomic Regulation Visual Biomarkers Closed-Loop AI Real-Time Detection Non-Invasive Smooth Pursuit Parasympathetic Activation Nervous System Balance Signal-Driven Self-Regulation

About the Project

Turning a common human failure mode into measurable signals and safer interventions

We’re building a research-backed, privacy-respecting approach to detect the onset of hijack states and help people regain control, especially when language arrives too late.

Early Stage and Recruiting Collaborators

How Does it Work?

3

Research pillars

  • Individuals + clinicians + researchers

We’re opening a collaboration path for university labs, clinicians, and applied researchers.

Get Updates
  • Signal-first approach
  • Evidence & validation
  • Privacy by design
  • Human autonomy focus

Key Research Fields

What we’re building around the amygdala hijack moment

We are exploring measurable signals that precede hijack states, and how AI can support early warning + de-escalation. Our aim is prevention-first, with careful validation and collaboration.

Autonomic & ocular signals

Exploring non-invasive markers (timing, attention, stress signatures) that can be measured safely.

State detection models

Developing models that infer risk states from signals—without labeling or judging people.

Intervention design

Designing gentle, user-controlled “brakes” (breath cues, pacing, prompts) that restore autonomy.

Validation & ethics

Building protocols for evaluation, bias checks, privacy safeguards, and clinician collaboration.

Why this matters

AI can accelerate decision velocity by removing friction, pre-selecting actions, and compressing the time humans have for reflection before acting.

When autonomic signals escalate, the nervous system can override deliberate choice in seconds. We’re building an AI assisted early warning layer that helps people notice the shift earlier and supports recovery of autonomy, without diagnosing, labeling, or replacing clinicians or personal judgment.

Human-first boundaries

We will not diagnose, label, or “score” people. The goal is awareness, warning, and safer choices.

Research collaboration

We are seeking partnerships with university labs and medical professionals to validate methods responsibly.

What we’re building

A prevention-first system: detect • warn • help restore control

We’re translating neuroscience into practical, measurable signals. The goal is to reduce “autonomy loss” moments in everyday life—work, parenting, conflict, and high-stakes situations—through early detection and user-controlled interventions.

1st

Primary focus

Catch the hijack earlier than language, then slow the cascade before decisions become irreversible.

Signal discovery

Identify candidate signals that correlate with acute stress shifts (non-invasive, privacy-aware, clinically relevant).

Protocol-first research

Define hypotheses, datasets, and validation methods with clinicians and researchers before building product claims.

Modeling & evaluation

Develop transparent evaluation: false positives/negatives, bias checks, and “when not to use” boundaries.

Open collaboration

We invite university professors and medical professionals to help steer research and validation.

Research Notes

Early hypotheses we’re exploring (public drafts)

How it works

From signals to safeguards — step by step, with validation

We start with observation, then move to modeling, then careful testing. We publish what we can, partner where we must, and keep claims conservative until evidence supports them.

Step 01

Signal discovery

Identify candidate non-invasive signals linked to stress shifts (attention, timing, autonomic proxies).

Step 02

Model & threshold design

Design models that detect state transitions and define safe thresholds (minimize false alarms).

Step 03

Validation with partners

Work with medical professionals and researchers to validate methods and define “do not use” boundaries.

Get involved

Choose how you want to participate (early access)

We’re pre-launch. Join the waitlist for early access to the first prototype, or subscribe for research updates. If you’re a university professor or medical professional, you can also collaborate with us.

Waitlist

Get early access when the first research-backed prototype is ready.

  • Early access emails
  • Private beta invitations
  • Progress snapshots
  • Feedback & pilot opportunities

Research Updates

Follow our research, experiments, and milestones as we build.

  • Monthly research summaries
  • Behind-the-scenes build notes
  • Neuroscience + AI insights
  • Announcements & public findings

Frequently Asked Questions

Clarifying what we mean and what we don’t

AmygdalaHijack.AI is an early-stage research initiative. These answers explain our intent, boundaries, and current focus.

An amygdala hijack refers to a rapid biological shift where threat-response circuitry can temporarily override calm reasoning and choice. People often recognize it afterward (“I snapped,” “I froze,” “I reacted without thinking”). Our work focuses on identifying the earliest measurable signals before that override fully takes hold.

No. AmygdalaHijack.AI is not a medical device, diagnostic tool, or provider of medical advice or treatment. We are conducting research into non-invasive, signal-based detection and user-controlled recovery support. Clinical decisions should always be made with qualified professionals.

We are starting with signal discovery and validation. That means identifying candidate indicators—such as timing shifts, attention changes, or autonomic proxies—and testing whether they reliably precede autonomy loss with minimal false alarms. Recovery support comes later, only after detection is well understood.

Yes. We are seeking collaboration with university researchers, clinicians, and applied neuroscience teams. Partners help define ethical boundaries, validation protocols, bias checks, privacy safeguards, and clear “do not use” conditions. Collaboration ensures this work remains responsible and evidence-driven.

You can join the waitlist or subscribe for updates using the form below. Subscribers receive concise research notes, milestone updates, and invitations to pilots when appropriate. We prioritize signal over noise and keep communication intentional.

Our Principles

Non-negotiables for building autonomy tech

Humans stay in control. We do not diagnose or label. We surface early signals, offer optional pauses, and design interventions that preserve choice—especially under speed and overload.

Principle 1

Autonomy over automation

We earn claims through measurement. We prioritize validated signals, publish what we can, and treat false alarms and bias as first-class problems. No impact claims without careful evaluation.

Principle 2

Proof over promise

Biological data is sacred. We minimize collection, keep processing local when possible, encrypt what must be stored, and give users control over retention. No dark patterns, no data resale.

Principle 3

Data minimization by design

Updates

Progress notes as we build AmygdalaHijack.AI