{ "source": "Cooked Index — occupational AI risk register", "page": "https://cookedindex.com/jobs/psychologists-all-other/", "methodology": "https://cookedindex.com/methodology", "notice": "Verdicts are re-examined as evidence accumulates. Re-fetch before relying on this; the page above always carries the current score.", "scored_at": "2026-08-11", "model": "claude-opus-5", "occupation": { "title": "Psychologists, All Other", "soc_code": "19-3039", "category": "Science", "us_employment": 18820, "median_annual_wage": 110840 }, "verdict": "EXPOSED", "risk_resistance": 54, "contested": false, "near_boundary": false, "dimensions": { "task_resistance": 11, "embodiment": 6, "liability_shield": 10, "trust_premium": 13, "judgment_accountability": 14 }, "reasoning": { "task_resistance": "At 11 the split is roughly even: the literature synthesis, scale scoring, norm comparison, regression on program-evaluation data, and boilerplate sections of a psychological report are already machine work, while the structured clinical interview, behavioral observation of a child or a police officer under stress, and cross-examination-proof formulation of findings still require a person in the room.", "embodiment": "A 6 reflects travel to jails, schools, clinics, training facilities and courtrooms plus hands-on test administration — WAIS blocks, gait or exertion protocols, developmental play tasks — but the environments are booked rooms with controlled conditions, and most of the week is still spent at a desk writing.", "liability_shield": "10 is the honest average of a mixed group: forensic and health psychologists in this residual bucket typically hold state licences and their signature on a fitness-for-duty or risk assessment is personally discoverable, but sport consultants, program evaluators, and PhD research psychologists in this same SOC code often practise under titles that no state licensing board controls.", "trust_premium": "13 sits at the point where relationship becomes the deliverable: a judge, a general counsel, a team's coaching staff, or a school district retains a specific named psychologist whose prior testimony and reputation they trust, and referral streams follow the individual rather than the practice — but unlike a therapist's caseload, most engagements are episodic and end when the report is filed.", "judgment_accountability": "14 is earned by the calls with no clean answer and real consequences — recommending a custody arrangement, declaring an officer unfit to return to duty, estimating violence risk when the PAI and the collateral records disagree, deciding whether a validity scale means malingering — where the psychologist names one conclusion and defends it under oath." }, "rationale": "This residual category bundles forensic, health, sport, developmental, and research psychologists — the modal worker sits in government, hospital, university, or consulting settings doing assessment, program evaluation, research design, and report writing rather than a full-time therapy caseload. Literature review, survey and test-score analysis, protocol drafting, and first-draft evaluation reports are now cheap for AI, which hollows out a large share of billable hours. What holds is the in-person clinical interview and behavioral observation, expert testimony and defensible sign-off on high-stakes determinations (custody, fitness-for-duty, risk of violence), and ownership of interpretive calls where instruments disagree — though only the licensed subset carries a real regulatory shield, and that shield is a policy choice, not physics.", "outlook": "Report and analysis hours compress sharply over the next decade while demand for licensed humans who conduct in-person assessments and defend high-stakes opinions holds steady, so the occupation splits between a shrinking research-writing tier and a durable licensed-determination tier.", "what_would_raise_it": { "levers": [ { "dimension": "liability_shield", "change": "State psychology boards or court rules explicitly barring AI-generated psychological evaluations from being admitted or billed unless a licensed psychologist personally conducted the interview and signs under personal liability — the model exists in APA's 2024-25 health-advisory work on AI and in state telehealth/AI-disclosure bills (e.g., Illinois' 2025 law restricting AI in therapy services and Utah's mental-health chatbot act). Also: forensic sign-off requirements under Daubert/FRE 702 as amended Dec 2023, where a testifying expert must vouch for the basis of the opinion.", "plausibility": "already happening", "would_add": 5 }, { "dimension": "judgment_accountability", "change": "Court-appointed or agency-appointed evaluator rosters (custody, fitness-for-duty, competency, violence risk under Tarasoff-type duties) naming an individual psychologist as the accountable decision-maker, with malpractice and board exposure attached to the determination rather than to the tool output.", "plausibility": "plausible", "would_add": 3 }, { "dimension": "task_resistance", "change": "Task-mix shift: once literature review, scoring, and first-draft reports are automated, the residual job is discordant-instrument interpretation, live behavioral observation, malingering/effort detection, and cross-examination-survivable reasoning. This occupation genuinely has two tiers, and the remaining tier is the judgment tier.", "plausibility": "already happening", "would_add": 3 }, { "dimension": "trust_premium", "change": "Institutional buyers (courts, insurers, professional sports leagues, disability adjudicators) writing contract language requiring a named human evaluator — e.g., SSA consultative examination rules or league concussion/mental-health protocols specifying an in-person licensed examiner.", "plausibility": "plausible", "would_add": 2 } ], "ceiling_note": "The research, program-evaluation, and psychometric-analysis segment of this residual category has no licensure hook and no named-human buyer; levers above apply almost entirely to the forensic and clinical-assessment subset, so the aggregate score for the SOC cannot rise far." }, "adjudication": null, "employment_history": { "points": [ { "y": 2017, "emp": 12880, "wage": 97740 }, { "y": 2018, "emp": 13480, "wage": 100770 }, { "y": 2019, "emp": 14220, "wage": 101790 }, { "y": 2020, "emp": 14960, "wage": 105780 }, { "y": 2021, "emp": 13800, "wage": 102900 }, { "y": 2022, "emp": 13900, "wage": 106420 }, { "y": 2023, "emp": 14480, "wage": 117750 }, { "y": 2024, "emp": 17790, "wage": 117580 }, { "y": 2025, "emp": 18820, "wage": 110840 } ], "from": 2017, "to": 2025, "change_pct": 46.1, "comparable_from": 2019, "spans_soc_revision": true }, "pivots": [ { "slug": "clinical-and-counseling-psychologists", "title": "Clinical and Counseling Psychologists", "verdict": "SAFE", "risk_resistance": 71, "median_wage": 100580, "overlap": 91, "skills_to_close": [ "Reading Comprehension" ] }, { "slug": "occupational-therapists", "title": "Occupational Therapists", "verdict": "SAFE", "risk_resistance": 79, "median_wage": 100330, "overlap": 85, "skills_to_close": [ "Operation and Control", "Instructing", "Operations Analysis", "Learning Strategies" ] }, { "slug": "physical-therapists", "title": "Physical Therapists", "verdict": "SAFE", "risk_resistance": 86, "median_wage": 102760, "overlap": 81, "skills_to_close": [ "Operations Analysis", "Operation and Control", "Equipment Selection", "Troubleshooting" ] } ], "license": "https://cookedindex.com/terms" }