Evidence-based. Clinician-reviewed.
Honest about what AI can and cannot do.
Every article on this blog is written against a clinical brief, grounded in peer-reviewed research, and reviewed by a licensed psychologist before publication. This page explains exactly how that process works.
Emotrek Clinical Team
Reviewed by a licensed clinical psychologist
Every Emotrek blog article is reviewed by a licensed clinical psychologist before publication. The reviewer's name and credentials will be published here once consent for public attribution is finalized. Until then, all reviewed content is published under the Emotrek Clinical Team byline. If you have questions about clinical accuracy or want to suggest a correction, contact us at clinical@emotrek.com.
Our evidence base
All factual claims in Emotrek content are sourced from peer-reviewed literature, primarily via PubMed. Citations are formatted as Author et al., Year (PMID: XXXXXXXX) and linked directly to the PubMed record. We do not cite opinion pieces, grey literature, or secondary sources for primary factual claims.
Where Israeli normative data exists for clinical instruments (e.g., Hebrew-validated PHQ-9), Hebrew-language articles reference it specifically. Where it does not exist, articles say so explicitly rather than assuming US norms apply.
Content review process
- 1
AI drafts from a clinical brief
Each article starts from a keyword and category brief. The AI is instructed to write peer-to-peer (therapist to therapist), cite PubMed, define every term, and answer the search query in the first paragraph.
- 2
Automated citation verification
Every cited PMID is verified against the PubMed eUtils API before the article reaches review. Articles with unverified citations are blocked from publication.
- 3
Clinical review
The article is reviewed by our licensed psychologist for factual accuracy, appropriate framing, and clinical safety. Articles targeting Israeli therapists are adapted for Israeli clinical context — not just translated.
- 4
Publication and freshness
Published articles are monitored for ranking decay. Articles where statistics or guidelines have been superseded are flagged for refresh and re-reviewed before the updated version is published.
AI ethics in clinical content
AI writes the first draft. Clinicians approve the final version. No article is published without human clinical review. The AI is a writing tool, not a clinical authority.
We do not use AI to generate clinical opinions. The AI follows a strict brief: cite primary sources, define terms, answer the search query directly. Opinions (“weekly PHQ-9 is overkill for low-acuity cases”) are drawn from cited literature, not generated.
AI summaries in the product are decision-support, not clinical decisions. Emotrek's pre-session summaries synthesize patient check-in data to help therapists orient before sessions. They do not diagnose, recommend treatment changes, or replace clinical judgment. The therapist is always the clinician.
Patient data is never used to train AI models. Mood entries, assessment scores, and session notes remain in the therapist's Firebase project. We do not aggregate or use patient data for model training.
Questions about our clinical content?
If you find a factual error, a citation that does not resolve, or content that feels clinically inappropriate, please contact us at clinical@emotrek.com. We take accuracy seriously and will respond within 48 hours.