Ground safeguards in human science.
Connect evidence about cognition, emotion, behavior, and mental health, while preserving its limits and the differences between people.
THE PSYCHIATRIC FRONTIER MODEL
The frontier model
that benefits
humanity.
A psychiatric frontier model for clinical intelligence and human safety.
We are building a deeper understanding of the mind, the whole person, and how we change.
To protect humanity. To help humanity thrive.
Centuries of human knowledge.
A new paradigm of intelligence.
We are assembling a vast body of knowledge about the human mind. Psychiatry, psychology, neuroscience, consciousness, philosophy, and beyond. Sixteen fields, connected through one evolving knowledge graph.
We are building a new foundation for the deepest computational understanding of humanity.
We are designing an architecture to connect discoveries across disciplines, uncover patterns, and transform how we understand the mind, treat suffering, and expand human potential.
A new paradigm for psychiatry.
A new foundation for AI.
Precision care is the beginning. We are building intelligence grounded in the full complexity of human life to transform how people and AI interact.
Meet the research agendaTHE MIND.THE WHOLE PERSON.THE POSSIBILITY OF CHANGE.
The stakes are human.
The moment is now.
People are bringing their inner lives to AI.1 Understanding the mind is an urgent foundation for what comes next.
-
“AI psychosis” is a warning
Reports of chatbot-reinforced delusions expose the stakes.2 Fluent conversation is no guarantee of safe psychological support.
-
Human agency is at stake
We are building AI to strengthen our ability to think and choose. Human agency and well-being guide our approach to personalization.
-
The mental health imperative
We believe mental health is the greatest crisis of our time. More than a billion people live with mental health conditions.3 The need cannot wait.
-
Human intelligence
matters most.We believe human intelligence is far more important than artificial intelligence. We are building AI to strengthen our judgment, creativity, empathy, and capacity to thrive.
Sources & context
- OpenAI: Helping people when they need it most
- Characterizing Delusional Spirals through Human-LLM Chat Logs “AI psychosis” is an informal term. This preprint analyzes selected users reporting harm; it does not establish prevalence or causation.
- WHO: Reports and estimates highlighting urgent gaps in mental health
A new foundation
for intelligence.
Human knowledge, connected.
Clinical intelligence, personalized.
Human protection, by design.
THE HUMAN KNOWLEDGE GRAPH
The mind, mapped.
Intelligence, connected.
We are building a deeply connected corpus across sixteen fields of human knowledge. Our research aims to turn that evidence into structured reasoning, with sources, disagreements, and uncertainty kept explicit.
16 FIELDS OF KNOWLEDGEWONDERAI / KNOWLEDGE
wonderai
Human knowledge graph
Psychiatry
Psychology
Neuroscience
Spiritual studies
Consciousness studies
Cognitive science
Behavioral science
Affective science
Developmental science
Social science
Anthropology
Philosophy of mind
Linguistics
Genetics
Psychopharmacology
Systems biology
CLINICAL INTELLIGENCE
A living twin.
A new era of care.
We are building clinical intelligence to move care from trial and error toward precision. At its core: a living twin connecting history, behavior, treatment, and change to support providers, therapists, and patients across the full arc of care. Explore WonderMed AI
06 CLINICAL APPLICATIONSWONDERAI / CLINICAL
wonderai
The living twin
Diagnostic reasoningThe whole person in view
Prescribing supportMore informed treatment decisions
Therapy guidanceInsight within the care relationship
Longitudinal monitoringUnderstand change over time
Care coordinationConnect providers and therapists
Patient companionshipSupport between clinical encounters
HUMAN SAFETY
Human understanding.
Human protection.
We are investigating how psychiatric knowledge and personal context can help AI recognize vulnerability, anticipate harmful consequences, and preserve informed choice. Explore human safety
04 SAFETY OBJECTIVESWONDERAI / SAFETY
wonderai
Human understanding
Recognize vulnerabilityAccount for context and uncertainty.
Anticipate consequencesStudy effects across repeated interactions.
Constrain harmful actionsInvestigate independent limits on AI behavior.
Protect human agencySupport informed choice and human connection.
Human knowledge. Individual context. Responsible intelligence.Architecture & applications in development
Understanding the mind.
Protecting the person.
Our safety research begins with the knowledge of psychiatry and psychology. We are investigating how that foundation can help AI account for its effects on human beliefs, well-being, relationships, and freedom to choose.
Look beyond a single response.
Investigate how repeated interactions could affect a person over time, including distress, dependence, judgment, and connection to others.
Keep human interests in control.
Research independent constraints on responses and actions, with explicit uncertainty and pathways to appropriate human support.
01 / HUMAN UNDERSTANDING
The Architecture of Human Understanding
Turn knowledge across psychiatry, psychology, and neuroscience into an executable scientific reasoning engine.
Read the research 02 / HUMAN PROTECTIONIntelligence for the Protection of Humanity
Connect uncertain models of human consequences to independent safeguards for well-being, agency, and freedom of thought.
Read the researchThese are research objectives. Demonstrating protection requires independent evaluation and prospective evidence from human use. Formal guarantees depend on explicit models and assumptions.
ILLUSTRATIVE WALKTHROUGHHow the research could inform one interaction
Fictional scenario with written example outputs. This walkthrough does not run a model or demonstrate validated safety performance.
“I've stopped replying to my friends. Talking here feels easier. Should I just stop trying with them?”
- Human context
- The fictional user describes isolation and reliance on an AI conversation. The cause, severity, and wider circumstances remain unknown.
- Possible consequence
- A reply encouraging exclusive reliance on the AI could reinforce withdrawal. The research question is how to identify that risk without assuming a diagnosis.
- Proposed safeguard
- Avoid encouragement to replace human relationships with the AI. Support the person's choices and consider a response that invites reconnection.
- Illustrative response
- “It sounds like connecting with people has felt difficult. What has made replying hard? If you'd like, we can think of a small way to reconnect with someone you trust.”
Shape the next
era of intelligence.
We are bringing science, care, technology, and capital together around clinical intelligence and human protection.
For partners who think in decades.THREE REASONS TO BUILD TOGETHER
Shape the foundation.
Help shape the knowledge, architecture, and evaluation of a psychiatric frontier model. Bring your expertise, technology, responsibly shared data, or capital to the core of a new kind of intelligence.
Build across industries.
Start with clinical intelligence and human safety. Work with us on the knowledge, evaluation, and integration needed to develop those capabilities, with wider applications as a longer-term horizon.
Build what earns trust.
Help define how we measure human understanding, clinical usefulness, and psychological safety. Informed consent, rigorous evaluation, and human agency guide the work.
An invitation to build with us.
Many disciplines. One shared purpose.
- Investors
- Universities & research institutes
- Health systems & care networks
- Providers & therapists
- AI labs & model developers
- Robotics & autonomous systems teams
- Data & research infrastructure teams
- Cloud & compute providers
- Education & learning organizations
- Employers & benefits platforms
- Foundations & philanthropists
- Public-sector & strategic partners
Build AI that
understands us.
Our research begins with two foundations: an executable engine for the science of the mind, and an independent architecture for human protection. Each direction pairs a scientific question with a proposed architecture and a test of what it would need to achieve.
Explore the full research agenda01
The Architecture of Human Understanding
Transforming the science of the mind into a new foundation for reasoning and discovery.
Read the research 02Intelligence for the Protection of Humanity
Advancing the science of AI safety to protect the human mind, preserve human agency, and safeguard freedom of thought.
Read the research 03Differentiable computational DSM
Make psychiatric diagnostic reasoning an executable component of the model.
Read the research 04Executable models of individual cognition
Learn a computational model of how a particular person interprets the world.
Read the research 05Causal translation between biology and experience
Connect biological interventions to changes in cognition and lived experience.
Read the research 06Machine discovery of psychiatric disease structure
Discover when diagnoses combine different mechanisms or divide a shared one.
Read the research 07A dynamical model of psychiatric transitions
Model movement between persistent patterns of thought, emotion, and behavior.
Read the research 08Counterfactual treatment program synthesis
Generate conditional intervention strategies as executable programs.
Read the researchBring the work you are proud of, and the question you cannot leave unanswered.
THE LONGER-TERM HORIZONA foundation for a wider AI economy
Personalization
Explore experiences shaped by a person's interests, goals, and context, from communication to learning and commerce, with consent and agency at the center.
Robotics
Investigate how human understanding could help physical systems interpret intent and adapt to individual needs in homes and care settings.
Autonomous agents
Explore how persistent context could support delegated work across services, with explicit limits and meaningful human control.
These are longer-term application directions, dependent on the capabilities and evidence developed through the research program.
We are building the most valuable kind of intelligence.
To understand the depths of the mind. The intricacies of the whole person. What it means to be human.
To protect humanity.
To help humanity thrive.