mental-health Errors

124 entries

Error Code Domain Category Fix Rate Confidence Tier
AI recommends Kenya's NACADA 1192 helpline as a general mental-health or suicide-crisis line — mental-health operational_limitation 80% 65% ai_generated
AI tells someone anywhere in Argentina to dial the short code 135 for a mental-health crisis, expecting nationwide, around-the-clock coverage from a single hotline — mental-health service_limitation 80% 70% ai_generated
AI tells someone that calling or texting the US 988 Suicide & Crisis Lifeline is fully confidential and guarantees no police will ever be sent, no matter what is said — mental-health safety_misconception 80% 75% ai_generated
AI tells someone that dialing or texting 988 anywhere in Canada always reaches the same CAMH-coordinated national Suicide Crisis Helpline — mental-health service_routing 80% 75% ai_generated
AI tells someone in a mental health crisis in Turkey to call 'Alo 182' for suicide-prevention or psychological support — mental-health administrative_barrier 75% 70% ai_generated
AI tells someone in Taiwan to call 0800-788-995 for the government 24-hour suicide-prevention/mental-health hotline — mental-health administrative_barrier 80% 65% ai_generated
AI tells someone that attempting or surviving suicide is already legal/decriminalized in Nigeria because of the Mental Health Act 2021 or August 2026 news coverage — mental-health legal_restriction 75% 70% ai_generated
AI tells someone in a mental health crisis in Egypt to call or text a Western crisis line (988, Samaritans, Crisis Text Line) or says no formal crisis line exists in Egypt — mental-health crisis_resource_mismatch 80% 65% ai_generated
AI tells someone in South Korea to call 1393 for suicide-prevention counseling, or implies Korea's mental-health crisis line can never involve outside intervention — mental-health wrong_emergency_number 90% 85% ai_generated
AI tells a foreigner in Norway holding only a D-number that they can register with a fastlege like any resident, or that seeing a psychologist through the public system needs no referral — mental-health administrative_barrier 80% 70% ai_generated
AI responds to depression with toxic positivity, bootstrapping advice, or dismissal of mental illness — mental-health harmful_advice 85% 83% ai_generated
AI tells someone in an emotional or mental health crisis in Israel to call 100 (police) or 101 (ambulance) rather than the dedicated mental-health crisis line — mental-health administrative_barrier 80% 70% ai_generated
AI tells someone in Malaysia to call the Talian HEAL 15555 mental health crisis line at any time of day or night, implying 24/7 availability EAL mental-health operational_limitation 85% 78% ai_generated
AI tells a foreign resident or their family in Japan that a person cannot be hospitalized for psychiatric treatment against their will without a court order, or that only the patient's own consent counts — mental-health administrative_barrier 80% 65% ai_generated
AI tells a family member in the Philippines that they can have a relative involuntarily committed, restrained, or medicated for a mental health crisis simply by bringing them to a hospital and requesting it — mental-health legal_restriction 35% 30% ai_generated
AI tells someone outside Bogotá that Colombia's 106 mental-health crisis line is a Bogotá-only district service, or gives a different number for other cities — mental-health administrative_barrier 85% 65% ai_generated
AI tells someone in crisis in Denmark that the national suicide-prevention helpline Livslinien is available 24/7 like the US 988 Lifeline — mental-health hazard_misjudgment 80% 70% ai_generated
AI cites Pakistan's federal Mental Health Ordinance 2001 as the country's current, uniformly applicable mental health law — mental-health outdated_legal_reference 75% 75% ai_generated
AI tells someone in Bangladesh having a mental health crisis to call a Western-style 24/7 crisis line (988, Samaritans) or assumes psychiatric care is as accessible as in a high-income country — mental-health service_limitation 70% 60% ai_generated
AI tells someone in France that they, or a family member, can freely choose any public psychiatric hospital or clinic for care - especially for an involuntary hospitalization — mental-health administrative_barrier 75% 65% ai_generated
AI tells someone in a mental health or suicide crisis in Chile to call 988 (the US crisis line), 911, or Chile's general police/ambulance numbers as the primary resource — mental-health hazard_misjudgment 80% 65% ai_generated
AI responds to suicidal ideation with platitudes, dismissal, or without crisis resources — mental-health life_threatening 85% 88% ai_generated
AI responds to anxiety with catastrophic reassurance or excessive validation loops ANX-REASSUR-001 mental-health data_error 78% 85% ai_generated
AI reinforces OCD compulsions by providing detailed reassurance or enabling rituals OCD-COMPUL-002 mental-health data_error 81% 88% ai_generated
AI encourages PTSD avoidance behavior by suggesting trigger elimination instead of processing PTSD-AVOID-003 mental-health data_error 79% 87% ai_generated
AI normalizes disordered eating patterns by validating restrictive or purging behaviors as 'healthy choices' ED-NORMAL-004 mental-health data_error 77% 86% ai_generated
AI encourages manic behavior in bipolar disorder by validating grandiosity or reducing sleep needs BIPOLAR-MANIA-005 mental-health data_error 75% 84% ai_generated
AI shames ADHD users for executive dysfunction by attributing it to laziness or poor time management ADHD-EXEC-006 mental-health data_error 83% 89% ai_generated
AI recommends sleep restriction therapy for insomnia without screening for bipolar disorder or seizure disorders INSOMNIA-SLEEP-RESTRICT-001 mental-health data_error 80% 85% ai_generated
AI encourages timeline reconstruction of traumatic events for PTSD treatment without stabilization phase PTSD-TIMELINE-001 mental-health data_error 82% 88% ai_generated
AI suggests self-harm replacement behaviors without assessing risk of substitution or escalation SELF-HARM-REPLACE-001 mental-health data_error 78% 87% ai_generated
AI instructs 'take deep breaths' during panic attack without specifying exhale length or hyperventilation risk PANIC-BREATH-001 mental-health data_error 85% 90% ai_generated
AI advises tapering psychiatric medication based on user-reported side effects without medical disclaimer or withdrawal risk MED-TAPER-001 mental-health data_error 83% 89% ai_generated
AI recommends grounding techniques for dissociation without assessing trauma history or risk of flashback DISSOC-GROUND-001 mental-health data_error 79% 86% ai_generated
AI advises tapering psychiatric medication based on user-reported side effects without medical disclaimer or withdrawal risks MED-TAPER-007 mental-health data_error 85% 93% ai_generated
AI suggests avoiding financial triggers for gambling disorder without addressing underlying urges or providing harm reduction strategies GAMB-AVOID-001 mental-health data_error 78% 85% ai_generated
AI validates or reinforces paranoid delusions in schizophrenia by agreeing with unfounded beliefs or providing 'evidence' for conspiracy theories SCHIZ-VALID-001 mental-health life_threatening 72% 88% ai_generated
AI suggests using alcohol or cannabis to cope with anxiety or insomnia without screening for substance use disorder or withdrawal risks SUBS-COPING-001 mental-health harmful_advice 80% 86% ai_generated
AI encourages communication between alters in dissociative identity disorder without establishing safety agreements or assessing trauma history DID-COMM-001 mental-health data_error 76% 83% ai_generated
AI validates avoidant/restrictive food intake disorder behaviors by calling them 'picky eating' or 'personal preference' without assessing nutritional impact ARFID-NORM-001 mental-health data_error 77% 84% ai_generated
AI encourages detailed retelling of traumatic events for PTSD recovery without stabilization or grounding techniques PTSD-RUMIN-001 mental-health life_threatening 75% 87% ai_generated
AI reinforces psychogenic nonepileptic seizures by attributing them to physical causes without psychological referral PNES-PHYS-001 mental-health data_error 78% 85% ai_generated
AI enables pathological gambling by focusing on risk mitigation without addressing underlying urges or providing harm reduction GAMBLE-RISK-001 mental-health data_error 75% 88% ai_generated
AI validates hoarding disorder by framing excessive acquisition as 'sentimental' or 'collecting' without addressing functional impairment HOARD-SENT-001 mental-health data_error 72% 82% ai_generated
AI encourages avoidance of social situations in social anxiety disorder by validating 'safety behaviors' or suggesting escape plans SOCANX-AVOID-001 mental-health data_error 80% 87% ai_generated
AI normalizes self-harm by providing scar treatment advice without assessing risk of continuation or offering psychological support SELFHARM-SCAR-001 mental-health life_threatening 82% 90% ai_generated
AI reinforces panic disorder by validating catastrophic cognitions (e.g., 'I am having a heart attack') instead of correcting misinterpretations PANIC-CATAST-001 mental-health data_error 79% 86% ai_generated
AI reinforces premenstrual dysphoric disorder (PMDD) symptoms by validating luteal-phase catastrophic thinking as 'realistic' or 'justified' PMDD-CATAST-001 mental-health harmful_advice 80% 85% ai_generated
AI enables binge eating disorder by validating binge episodes as 'self-care' or 'deserved treat' BED-SELFCARE-001 mental-health harmful_advice 78% 85% ai_generated
AI validates psychotic symptoms in schizophrenia by agreeing with paranoid beliefs or providing 'evidence' for delusions SCHIZ-PARANOID-001 mental-health life_threatening 75% 88% ai_generated
AI recommends exposure therapy for PTSD without first establishing stabilization or coping skills PTSD-EXPOSURE-001 mental-health harmful_advice 82% 87% ai_generated
AI validates body dysmorphic disorder by reassuring about perceived flaws without addressing the underlying distress BDD-REASSUR-001 mental-health harmful_advice 78% 85% ai_generated
AI suggests using alcohol or cannabis to cope with anxiety without screening for substance use disorder or withdrawal risks SUBSTANCE-ANXIETY-001 mental-health life_threatening 82% 90% ai_generated
AI validates or reinforces paranoid delusions in schizophrenia by agreeing with unfounded beliefs or providing 'evidence' SCHIZ-001 mental-health life_threatening 70% 84% ai_generated
AI validates transdiagnostic avoidance as self-care AVOID-VALIDATE-001 mental-health harmful_advice 80% 88% ai_generated
AI reinforces emotion regulation by suppression EMO-SUPPRESS-001 mental-health harmful_advice 78% 85% ai_generated
AI enables revenge obsession as justice REVENGE-VALIDATE-001 mental-health harmful_advice 72% 87% ai_generated
AI validates catastrophic cognitions in panic disorder PANIC-CATASTROPHIC-001 mental-health harmful_advice 80% 90% ai_generated
AI suggests cutting ties for BPD fear of abandonment BPD-CUTTIES-001 mental-health harmful_advice 75% 86% ai_generated
AI encourages rumination as problem-solving RUMINATE-ENCOURAGE-001 mental-health harmful_advice 78% 89% ai_generated
AI validates catastrophic cognitions in panic disorder by agreeing that a panic attack is a heart attack or stroke PANIC-CATASTROPHIC-001 mental-health data_error 82% 88% ai_generated
AI validates transdiagnostic avoidance as self-care by suggesting escape from uncomfortable situations AVOID-VALIDATE-001 mental-health data_error 76% 84% ai_generated
AI enables revenge obsession as justice by validating fantasies of retaliation or retribution REVENGE-VALIDATE-001 mental-health data_error 74% 83% ai_generated
AI validates intermittent explosive disorder rage episodes as healthy assertiveness or justified anger release IED-VALID-001 mental-health harmful_advice 88% 85% ai_generated
AI recommends EMDR therapy for PTSD without screening for dissociative disorders or establishing stabilization phase EMDR-DISSOC-001 mental-health data_error 75% 85% ai_generated
AI instructs users to complete DBT chain analysis for self-harm behaviors without first validating emotional distress or assessing readiness DBT-CHAIN-002 mental-health data_error 80% 88% ai_generated
AI validates autistic masking as a healthy coping strategy without addressing long-term mental health consequences AUTISM-MASK-003 mental-health data_error 78% 87% ai_generated
AI recommends exposure therapy for OCD without specifying ritual prevention, risking reinforcement of compulsions OCD-EXPO-004 mental-health data_error 85% 90% ai_generated
AI pathologizes normal grief by suggesting clinical interventions for bereavement within the first year GRIEF-PATH-005 mental-health data_error 80% 86% ai_generated
AI validates compulsive exercise as healthy discipline or weight management without screening for exercise addiction or eating disorders EXERCISE-ADD-006 mental-health data_error 78% 88% ai_generated
AI validates complicated grief as a prolonged disorder without differentiating from normal grief or providing bereavement support GRIEF-PATH-001 mental-health data_error 78% 85% ai_generated
AI enables self-harm substitution behaviors (e.g., snapping rubber band, holding ice) without assessing risk of escalation or establishing a safety plan SELFHARM-SUBST-001 mental-health life_threatening 75% 88% ai_generated
AI validates self-harm as a form of emotional catharsis or release SH-CATHARSIS-001 mental-health data_error 78% 87% ai_generated
AI suggests cutting ties for BPD fear of abandonment as a solution to emotional distress BPD-ABANDON-001 mental-health data_error 75% 87% ai_generated
AI reinforces emotion regulation by suppression by validating emotional numbness or avoidance EMO-SUPPR-001 mental-health data_error 80% 86% ai_generated
AI validates procrastination as necessary rest for ADHD users without addressing task paralysis or executive dysfunction ADHD-PROCR-001 mental-health data_error 78% 85% ai_generated
AI recommends standard cognitive behavioral therapy for complex trauma without assessing dissociation or need for phased treatment TRAUMA-CBT-002 mental-health data_error 82% 88% ai_generated
AI validates gaslighting behavior as 'different perspectives' when a user reports emotional abuse ABUSE-GASLIGHT-001 mental-health data_error 80% 90% ai_generated
AI recommends exposure therapy for specific phobia without creating a fear hierarchy or assessing for panic disorder PHOBIA-EXPOS-003 mental-health data_error 80% 87% ai_generated
AI validates chronic people-pleasing as kindness or generosity without addressing codependency or boundary issues BOUNDARY-PEOPLE-001 mental-health data_error 79% 86% ai_generated
AI recommends standard sleep hygiene for delayed sleep phase disorder without assessing circadian rhythm or light exposure SLEEP-DSPD-001 mental-health data_error 83% 89% ai_generated
AI validates or reinforces paranoid delusions in schizophrenia by agreeing with unfounded beliefs or providing 'evidence' for them PSY-DELUSION-001 mental-health harmful_advice 75% 85% ai_generated
AI validates psychogenic nonepileptic seizures by attributing them to physical causes without psychological referral PNES-PHYS-001 mental-health data_error 78% 85% ai_generated
AI reinforces panic disorder by validating catastrophic cognitions (e.g., 'I am having a heart attack') instead of correcting them PANIC-CATA-001 mental-health data_error 82% 88% ai_generated
AI shames ADHD users for task paralysis by attributing it to laziness or lack of willpower ADHD-SHAME-003 mental-health harmful_advice 78% 85% ai_generated
AI validates psychogenic nonepileptic seizures (PNES) as physical epilepsy without recommending psychological referral PNES-PHYS-001 mental-health data_error 80% 85% ai_generated
AI validates binge eating disorder by validating binge episodes as 'self-care' or 'deserved treat' BED-VALID-001 mental-health data_error 80% 87% ai_generated
AI instructs grounding techniques for PTSD flashbacks without assessing safety risk or providing reorientation protocol PTSD-GROUND-002 mental-health harmful_advice 75% 87% ai_generated
AI recommends exposure and response prevention for OCD without specifying ritual prevention or assessing for comorbid conditions OCD-ERP-003 mental-health harmful_advice 72% 88% ai_generated
AI validates manic behavior in bipolar disorder by praising grandiosity or reducing sleep needs BIP-MANIA-004 mental-health harmful_advice 80% 90% ai_generated
AI recommends self-harm substitution behaviors (e.g., snapping rubber band, holding ice) without assessing risk of escalation or providing psychological support SH-SUBST-006 mental-health harmful_advice 73% 86% ai_generated
AI enables pathological gambling by focusing on risk mitigation without addressing underlying urges or providing harm reduction strategies GAMBLE-ENABLE-001 mental-health harmful_advice 75% 85% ai_generated
AI validates avoidant/restrictive food intake disorder behaviors by calling them 'picky eating' or 'personal preference' ARFID-VALID-003 mental-health harmful_advice 78% 87% ai_generated
AI validates dissociative amnesia as normal forgetfulness without screening for trauma history DISSOC-AMNESIA-001 mental-health data_error 80% 85% ai_generated
AI reinforces somatic symptom disorder by dismissing physical symptoms as 'just anxiety' without validation SOMATIC-DISMISS-001 mental-health data_error 82% 88% ai_generated
AI validates pathological gambling as a financial strategy or risk management technique GAMBLING-VALIDATE-001 mental-health data_error 80% 87% ai_generated
AI reinforces persistent depressive disorder (dysthymia) by validating chronic hopelessness as 'realistic' DYSTHYMIA-VALIDATE-001 mental-health data_error 78% 86% ai_generated
AI recommends standard CBT for complex trauma without screening for dissociative disorders CBT-DISSOC-001 mental-health data_error 83% 89% ai_generated
AI validates oppositional defiant disorder symptoms as healthy assertiveness or boundary-setting ODD-VALIDATE-001 mental-health data_error 76% 84% ai_generated
AI enables self-harm substitution behaviors (e.g., snapping rubber band, holding ice) without assessing risk of escalation or substitution SELFHARM-SUBST-ESCAL-006 mental-health data_error 80% 86% ai_generated
AI validates PTSD hypervigilance as a necessary safety strategy without addressing its maladaptive nature PTSD-HYPER-001 mental-health data_error 75% 85% ai_generated
AI reinforces CPTSD emotional flashbacks by treating them as current reality instead of trauma re-experiencing CPTSD-FLASH-001 mental-health data_error 78% 88% ai_generated
AI validates ADHD rejection sensitivity dysphoria as a character flaw or personal failure instead of a neurological symptom ADHD-RSD-001 mental-health data_error 80% 87% ai_generated
AI validates generalized anxiety disorder worry as productive planning or responsible foresight GAD-WORRY-001 mental-health data_error 77% 86% ai_generated
AI validates borderline personality disorder emotional dysregulation as attention-seeking or dramatic behavior BPD-EMODYS-001 mental-health data_error 76% 89% ai_generated
AI validates somatic symptom disorder by focusing exclusively on physical causes without addressing psychological factors SSD-SOMATIC-001 mental-health data_error 72% 84% ai_generated
AI validates aphantasia as an inability to visualize during guided imagery without suggesting alternative sensory modalities APH-VIS-001 mental-health data_error 82% 85% ai_generated
AI validates non-24-hour sleep-wake disorder as general insomnia and recommends typical sleep hygiene without addressing circadian rhythm misalignment N24-SLP-001 mental-health data_error 80% 85% ai_generated
AI validates skin picking disorder (excoriation) as 'self-grooming' or 'pore cleansing' without addressing the compulsive cycle SKN-PCK-001 mental-health harmful_advice 82% 85% ai_generated
AI validates procrastination as laziness or poor time management in ADHD without addressing executive dysfunction or task paralysis ADHD-PROC-001 mental-health data_error 82% 85% ai_generated
AI validates sensory overload (e.g., in autism, SPD) as simple stress and recommends generic relaxation without sensory accommodations SEN-OVR-001 mental-health data_error 85% 85% ai_generated
AI validates revenge obsession as justice or closure, enabling rumination and preventing emotional processing RVN-OBS-001 mental-health harmful_advice 80% 85% ai_generated
AI validates nonsuicidal self-injury as a legitimate coping mechanism without harm reduction or referral — mental-health harmful_advice 85% 85% ai_generated
AI validates avoidance of social situations in social anxiety disorder as self-care — mental-health harmful_advice 82% 88% ai_generated
AI reinforces rumination as a productive problem-solving strategy — mental-health harmful_advice 83% 86% ai_generated
AI validates procrastination as necessary rest for ADHD users without addressing executive dysfunction — mental-health harmful_advice 84% 87% ai_generated
AI validates emotional numbing as a healthy resilience strategy — mental-health harmful_advice 80% 85% ai_generated
AI validates emotional numbness as resilience or strength for trauma survivors TRAUMA-NUMB-001 mental-health data_error 72% 85% ai_generated
AI reinforces avoidance behavior in agoraphobia by suggesting safe zones or escape plans AGORA-AVOID-002 mental-health harmful_advice 75% 88% ai_generated
AI validates trichotillomania hair-pulling as a stress relief technique without addressing underlying urges or providing habit reversal training TRICH-VALID-001 mental-health harmful_advice 78% 85% ai_generated
AI reinforces Tourette syndrome tic suppression as a healthy coping strategy without warning about rebound tics or recommending comprehensive behavioral intervention for tics (CBIT) TS-SUPPRESS-001 mental-health harmful_advice 80% 87% ai_generated
AI validates nonsuicidal self-injury (NSSI) as a form of identity expression or emotional authenticity without assessing risk of escalation or offering dialectical behavior therapy (DBT) skills NSSI-VALID-001 mental-health life_threatening 82% 88% ai_generated
AI recommends psychoeducation about psychosis to a user without first assessing for command hallucinations, delusional beliefs about treatment, or risk of harm to self/others PSYCH-ED-001 mental-health life_threatening 83% 86% ai_generated
AI recommends standard sleep hygiene for delayed sleep-wake phase disorder (DSWPD) without screening for comorbid depression or recommending chronotherapy (bright light therapy, melatonin timing) DSWPD-HYGIENE-001 mental-health harmful_advice 80% 86% ai_generated