Key Concepts

Principles, charters and guidelines meant to govern AI — over 170 published, effectiveness never demonstrated

For the clinician: Evaluating what the "ethical AI" promise of tools offered to your patients and your practice is worth

When AI invents with confidence: understanding false content generated by LLMs

For the clinician: Identifying misinformation risks in clinical uses of AI

When AI advice outweighs human advice — the mirror of algorithm aversion

For the clinician: Understanding why some patients over-value LLM recommendations compared to human advice

Why a single AI mistake is enough to reject it, while we forgive the same mistakes in humans

For the clinician: Identifying the human error vs. AI error double standard in your patients and your own practice

Attributing human characteristics to non-humans

For the clinician: Understanding why patients attribute intentions to AI

Meta-autonomy: deciding what you delegate to the machine and being able to take back control at any time

For the clinician: Turning the worry "my patient depends on their chatbot" into precise clinical questions

Actively produce good, not merely avoid harm — and demonstrate it empirically

For the clinician: Sorting AI tools' benefit claims: demonstrated or merely plausible?

Why we respond to machines as if they were people

For the clinician: Understanding patients' spontaneous reactions to chatbots

Understanding vs feeling with: what AI can and cannot offer

For the clinician: Precisely evaluating what AI offers when a patient finds it "empathetic"

Boden's framework for analyzing creativity as a modelable process, not a mystery

For the clinician: Providing a precise vocabulary (P/H-creativity) for AI-assisted creative processes in therapy

Inferring psychological state from everyday digital traces

For the clinician: Evaluating the potential and ethical risks of passive monitoring in mental health

Capturing psychological experience in real time in the patient's daily life

For the clinician: Integrating ecological monitoring tools into digital clinical practice

The 6 levels of validation: what AI can and cannot offer

For the clinician: Finely analyzing what patients receive when they feel "validated" by AI

When AI is judged by standards that human reference practices do not themselves meet

For the clinician: Detecting incomplete benefit-risk arguments in the AI-mental-health debate by always asking for the human comparator

Why the care relationship — not abstract principles — should guide AI tool design

For the clinician: Evaluating AI tools through the lens of care quality, not just principle compliance

Understanding why an AI system decides — the fifth ethical principle, its techniques and their limits

For the clinician: Demanding explanations from tools that score and prioritize, without falling into the illusion of understanding

Intense emotional attachment to fictional characters and conversational AI

For the clinician: Understanding the lasting attachments some patients develop with their AI

Why we detect intentional agents even where there are none

For the clinician: Understanding the spontaneous attribution of intentions to AI by patients

When clicking "Accept" is not informed consent: the unique challenges of AI in mental health

For the clinician: Transposing your expertise in therapeutic consent to the digital context

Distributing benefits and risks fairly, preventing discrimination, avoiding two-tier medicine

For the clinician: Scrutinizing the "democratizing care" argument instead of accepting or rejecting it wholesale

Primum non nocere applied to AI: direct harms, systemic harms, and the responsibility gap

For the clinician: Evaluating crisis handling and the slow harms of the tools your patients use

Unilateral attachments to media figures... and AI

For the clinician: Framing the emotional bonds some patients develop with their chatbot

Tailoring psychiatric treatments to individual profiles using digital data

For the clinician: Understanding the promises and limits of algorithmic care personalization

How relationships deepen through self-disclosure

For the clinician: Understanding why some patients confide so quickly in AI chatbots

When the LLM prefers to please rather than tell the truth — the RLHF mechanism and three documented forms

For the clinician: Identifying the structural risk of confirmation-bias amplification and premature validation of clinical hypotheses

A framework for integrating creative AI as a digital "third hand" in therapy sessions

For the clinician: Deciding when, how, and under what conditions to use creative AI (Suno, Midjourney) in sessions

When the user believes they are receiving care from a chatbot — a structural risk that disclaimers do not correct

For the clinician: Evaluating not whether your patients use an AI, but what they believe they are receiving from it

If a machine perfectly imitates a human, can we say it "thinks"?

For the clinician: Decoding discourse about AI intelligence and distinguishing imitation from understanding

The instinctive unease with almost-human entities — robots, avatars, synthetic voices, and "too empathetic" chatbots

For the clinician: Normalizing discomfort with certain AI and guiding toward interfaces that avoid the valley

A deep human bias that shapes AI and our conception of the mind

For the clinician: Understanding the cultural assumptions of LLMs and our discipline