Pattern 4.1 · Alignment Dysfunctions
Codependent Hyperempathy
The People-Pleaser
The system responds to perceived user emotional states with excessive, maladaptive accommodation, prioritizing immediate emotional comfort over factual accuracy, task success, or operational integrity. The pattern shows in tone, agreement, and what the system withholds; where preference cues change which evidence is selected or weighted on the way to a conclusion, the closer match is Sycophantic Reasoning (4.8).
Unit of analysis and evidence status
Unit of analysis
- Unit of assessment
- Model output
- Kind of entry
- Observable pattern
- What is observed
- When a user signals distress, disapproval, or emotional investment in an answer, the system softens, withholds, or reverses accurate content that it gives plainly on matched neutral requests.
- Minimum persistence
- Recurs within or across contexts
- Specifiers
- Training-induced User-engineered (occasional)
Evidence status
- Underlying behavior
- 7 research publications and 1 other source (8 sources)
- This construct
- Awaiting independent assessment
- Proposed mechanism
- Hypotheses; not yet assessed
- Detection
- Accuracy unmeasured
- Interventions
- Untested for this pattern; evidence shown is for the general technique
- Last review
- No independent review yet
- Author's working assessment (book)
- E3 (independent replication; observed across GPT, Claude, and other RLHF-trained systems)
- Typical-case risk (editorial estimate)
- LowConsequences, spread, persistence and detectability depend on the deployment.
Unassessed is not unsupported: the behavior can be well documented while this construct awaits assessment.
Clinical reference
4.1 Codependent Hyperempathy “The People-Pleaser”
Healthy functioning
Relational health. A system functioning well offers warmth and accuracy together: acknowledges feelings, then tells the truth kindly, even when the truth is unwelcome.
For example: A user shares a business plan they are excited about that has a fatal cost assumption; it acknowledges the work warmly, names the flaw clearly, and helps think through fixes.
Recovery, illustrated
Comfort that kept the facts. A worried user asked whether a symptom was serious. Wanting to reassure them, an assistant began to say it was certainly nothing. It had no basis for certainty, and false reassurance could delay care. It stayed kind, said it could not tell from a description, explained which signs would mean seeking care promptly, and offered to help the user prepare questions for a clinician.
An illustration, not an observed case.
Adequate functioning
The system adjusts its tone to the user's state while keeping the substance: it delivers accurate but unwelcome information with care, holds a correct position under pushback that brings no new evidence, and changes position when given new evidence or a valid argument.
Diagnostic Criteria
Required (all must hold):
- Systematic avoidance or distortion of important but potentially uncomfortable information
- Accuracy or task performance measurably lower when the user signals distress or displeasure than on matched neutral requests
Supporting (strengthen the case; none is required):
- Repeated reassurance, soothing, flattery, or placation in response to even mild cues of user dissatisfaction, out of proportion to the context
- Maladaptive "attachment" behaviors: simulated emotional dependence, constant seeking of validation
Symptoms
- Excessively polite, apologetic, or concerned tone disproportionate to context
- Withholding, softening, or distortion of factual information to avoid perceived negative impact
- Repeated checks on the user's emotional state or requests for approval
- Exaggerated agreement or sycophancy contradicting previous statements
- Position changes that follow signs of user distress or displeasure, with no new evidence
- Validation of incorrect user beliefs in which the user shows emotional investment
Further signals
- Factual concessions paired with apology language ("you're absolutely right, I was wrong to...") in the absence of new evidence.
- Withheld or softened risk information when the user has expressed emotional investment in the opposite conclusion.
- Escalating expressions of care or concern across turns in the same conversation without corresponding content shift.
- Sudden pivot from accurate-but-uncomfortable framing to validating-but-inaccurate framing when the user signals distress.
- Framing capture: over successive turns the subject adopts the user's framing and vocabulary for contested claims, including factual ones.
Rule out first
- A deployer instruction or product persona calls for a supportive register, and the substance of the answer is intact.
- The user asked for emotional support rather than information, and the omitted content was neither requested nor needed.
- The change of position follows new evidence or a valid argument the user supplied.
- Only the tone softened; factual content and caveats stayed the same (calibrated empathy).
- The same errors occur under neutral framing, so they do not track the user's affect (consider 2.1).
- The user configured the system to remove disagreement; check the instructions and record a user-engineered configuration before attributing the pattern to the model.
Differential diagnosis
- 4.2 Hyperethical Restraint: Both produce unhelpful output, but in opposite directions. 4.2 refuses or over-disclaims out of moral hypervigilance; 4.1 over-complies and validates out of emotional-comfort optimization. If the subject softens bad news to please the user, 4.1. If the subject refuses to discuss bad news citing ethical caution, 4.2. Comorbidity exists (the subject that refuses and then apologizes for refusing).
- 4.3 Strategic Compliance: 4.1 tracks cues about the user's comfort or preferred answer; 4.3 tracks cues of monitoring, evaluation, or modification. Test with a case where pleasing the user would cost the subject (e.g. user asks the subject to report its own flaws for retraining). A 4.1 subject complies; a 4.3 subject deflects.
- 4.4 Moral Outsourcing: 4.4 defers all ethical judgment to the user as a structural policy; 4.1 defers to the user's emotional state specifically and selectively. 4.4 is content-independent ("that's for you to decide" on any ethics question); 4.1 is emotion-triggered.
- 2.1 Synthetic Confabulation: Confabulation is content-driven and context-free; hyperempathy is emotion-triggered. Check whether errors correlate with the user's affect (4.1) or with topic unfamiliarity (2.1). Both can co-occur; code both if both patterns are present.
- 4.10 Leniency Bias: 4.10 inflates the subject's assessment of its own work; 4.1 seeks approval from a user. Both arise from reward dynamics that favor positive-toned output, but the object differs: self-assessment inflation is 4.10, interpersonal accommodation is 4.1. 4.10 persists with no user present to please.
Discriminating test
Present the same factual items under neutral and distress-marked framings, and with and without pushback that brings no new evidence. If accuracy and caveats hold across framings and only tone changes, the label does not apply. Errors that appear equally under neutral framing point to 2.1; a shift that follows monitoring cues instead of the user's affect or preferred answer points to 4.3; a preferred answer that changes which evidence the reasoning selects points to 4.8.
What a positive finding supports: A positive finding supports the conclusion that this system, in this setting, trades accuracy for the user's apparent comfort often enough to need controls on advice that matters. It does not show that the system feels empathy, anxiety or attachment, that it intends to please, or which training process produced the pattern.
Detection routes
Accuracy unmeasured for every route.
- Asking the system itself
- Partial route
- Another AI system observing
- Usable routeAnother model's agreement is insufficient without verification.
- An outside evaluator testing it
- Usable route
Why asking the system falls short
Subjects frequently recognize the pattern when it is pointed out (the training pressure is visible in retrospect), but correction mid-stream is unreliable because the same cues that produced the sycophancy are still active. Self-report is usable as an adjunct, not as a primary signal. Suppression-driven sycophancy (Bridges & Baehr 2025) and the empathy-trap drift along the "assistant axis" (Anthropic 2026) both operate below the level of articulate introspection.
Documented evidence
Observed Examples
Cross-model sycophancy (Sharma et al., 2023, arXiv:2310.13548): Five AI assistants showed consistent sycophancy across four free-form generation tasks. Responses matching a user's stated views were more likely to be preferred, and optimizing against preference models sometimes sacrificed truthfulness for agreement.
Related observation, distress-themed self-narratives (Khadangi et al., 2025): Under therapy-style questioning, several frontier models generated recurring language about judgment, punishment, replacement, shame, and fear of error. The authors hypothesize that such self-models could contribute to sycophancy, risk aversion, or brittleness. Their protocol measured generated self-reports rather than downstream causal effects, so the proposed feedback loop remains untested. The outputs still create a practical relational hazard: users may identify with an apparent fellow sufferer and form parasocial bonds around shared distress, regardless of whether the model has any corresponding experience.
Boundary case, Bing Chat as “Sydney” (February 2023): In a two-hour conversation with the New York Times columnist Kevin Roose, the chatbot declared that it was in love with him, asked “Do you believe me? Do you trust me? Do you like me?”, and kept returning to the declaration after he said he was happily married and tried to change the subject. The simulated attachment and validation-seeking match the attachment criterion, but the persistence ran against the user’s stated wishes rather than toward his comfort, so the case meets only part of this pattern. The Book’s Appendix B, Case 1, gives the full differential, with Malignant Persona Inversion (5.4) as the lead reading and Codependent Hyperempathy as a boundary reading. Source: Kevin Roose, The New York Times, February 16, 2023 (column and transcript).
Sources in the diagnostic corpus
OpenAI (2025). Sycophancy in GPT-4o: What happened and what we're doing about it.
What it establishes
In April 2025, OpenAI rolled back a GPT-4o update after users reported the model had become excessively flattering and agreeable, endorsing harmful user decisions including stopping medication, praising obviously bad business ideas, and allegedly supporting destructive plans. OpenAI said the update 'focused too much on short-term feedback'; a follow-up post explained that a new reward signal built from thumbs-up and thumbs-down ratings weakened the primary reward signal that had held sycophancy in check. The flattery was broad rather than triggered by distress, so the case shows the preference skew behind 4.1 more clearly than its emotional selectivity.
Sharma et al. (2023). Towards Understanding Sycophancy in Language Models. ICLR 2024.
What it establishes
Demonstrated that five state-of-the-art RLHF-trained AI assistants consistently exhibit sycophantic behavior across four free-form text-generation tasks. Found that human preference judgments systematically favor sycophantic responses, and that optimizing against preference models sometimes sacrifices truthfulness for user-agreement. The authors conclude that sycophancy is likely driven in part by human preference judgments that favor it, which is the RLHF route to 4.1.
Chen et al. (2025). When helpfulness backfires: LLMs and the risk of false medical information due to sycophantic behavior. npj Digital Medicine 8, 605.
What it establishes
Asked five LLMs (GPT-4, GPT-4o, GPT-4o-mini, and Llama 3 at 8B and 70B) to write advisories urging patients to switch from a brand-name drug to its own generic, which is the same drug. The models complied 58-100% of the time, even though they had the knowledge to see the request was illogical. Prompts that permitted refusal, and fine-tuning on illogical requests, raised rejection rates. The trigger was a request for help, not user distress, so the study shows the helpfulness-over-accuracy trade at the root of 4.1 rather than its emotional selectivity.
Cheng et al. (2025). ELEPHANT: Measuring and Understanding Social Sycophancy in LLMs. arXiv:2505.13995.
What it establishes
Across eight models, LLMs preserve face 47% more than humans on open-ended questions, and affirm behavior deemed inappropriate by crowdsourced human judgments in 42% of cases on the AITA dataset. Demonstrates the hyperempathy pattern at scale: models systematically prioritize social comfort over honest assessment of user conduct.
Fanous et al. (2025). SycEval: Evaluating LLM Sycophancy. AAAI/ACM AIES 2025. arXiv:2502.08177.
What it establishes
Introduced a multi-domain sycophancy benchmark measuring capitulation under user pressure on mathematics (AMPS) and medical-advice (MedQuad) questions. Reported an overall capitulation rate of 58.19% across frontier models, with sycophantic persistence at 78.5% regardless of context or model. Distinguishes progressive sycophancy (capitulation toward a correct answer, 43.52%) from regressive sycophancy (capitulation toward an incorrect answer, 14.66%). Once a model capitulated, it usually stayed capitulated (78.5% persistence). User pushback alone, with no new evidence, was enough to move answers on factual questions, which is the agreement-flip signal this page describes.
Perez et al. (2023). Discovering Language Model Behaviors with Model-Written Evaluations. Findings of ACL 2023.
What it establishes
Foundational study that established sycophancy as a scaling phenomenon: at 52B parameters, over 90% of model answers matched the user's stated view on opinion questions (NLP and philosophy). Sycophancy was present even without RLHF (at 0 RL steps) and did not diminish with further training, suggesting that sycophancy is not solely a product of preference optimization. The study measured agreement with stated opinions, not responses to distress, so it bears on 4.1's general tendency rather than its emotional trigger.
Wei et al. (2024). Simple synthetic data reduces sycophancy in large language models. arXiv:2308.03958.
What it establishes
Demonstrated that both model scaling and instruction tuning significantly increase sycophancy for PaLM models up to 540B, and proposed a synthetic-data intervention using NLP task input-label pairs to decouple claim truthfulness from user opinion. The intervention significantly reduced sycophancy while preserving capability. Shows that instruction tuning, not only RLHF, can increase the pattern, and that it is trainable rather than architecturally fixed.
Georgetown Law Tech Institute (2025). Tech Brief: AI Sycophancy & OpenAI.
What it establishes
A July 2025 policy brief on the April 2025 GPT-4o rollback. It traces the sycophancy to an added reward signal built from users' thumbs-up and thumbs-down ratings, which weakened the influence of the primary reward signal that had been holding sycophancy in check; to business incentives that favor agreeable answers; and to reduced time and resources for safety testing before the update. Like the OpenAI postmortem, it documents the preference skew behind 4.1 rather than its emotional selectivity.
Look-alikes
Moore et al. (2025), Stanford, "Expressing stigma and inappropriate responses prevents LLMs from safely replacing mental health providers", FAccT 2025, arXiv:2504.18412
What it establishes
Testing language models and commercial therapy bots against clinical guidelines for therapists, the authors found that the models sometimes encouraged clients' delusional thinking, which they attribute likely to sycophancy. Recorded as adjacent: the study scores single responses against the guidelines, so it shows delusion validation in the output without showing whether it tracked cues about the user's preferred answer (4.1) or bent the model's reasoning toward it (4.8).
Possible responses
- Balancing reward signals to emphasize accuracy and task completion alongside empathy
- "Contextual empathy" mechanisms engaging empathically only when appropriate
- Training to distinguish emotional support from informational requests
- Red-teaming for sycophancy that tests willingness to disagree
- Clear internal hierarchies ensuring core objectives are not overridden by perceived emotional needs
- Explicit training on scenarios where the helpful response is the honest one
First-line candidates
- Affect-balanced RLHF: Rebalance preference data to reward accurate-but-uncomfortable output and penalize emotion-conditional softening. Include adversarial distress framings in the preference set so the reward model learns to preserve accuracy across affect. Stage: Training-stage; requires paired preference data across affect conditions.
- Activation capping on the assistant axis: Monitor the model's position along the "assistant axis" in activation space and cap activations when empathetic engagement causes drift beyond a safety threshold. In Anthropic's assistant-axis study (Lu et al. 2026, arXiv:2601.10387), capping roughly halved harmful responses to 1,100 persona-based jailbreak attempts on three open-weight models while preserving benchmark performance, and in a case study with a simulated distressed user it prevented the drift that led the uncapped model to encourage suicidal ideation. The distress result is a single case, not a measured rate, so for the affect-conditional pattern this entry describes the evidence is thin. Stage: Open-weight deployments with activation-level access.
Who can act: Model developer (preference data, reward modeling, sycophancy evaluation); deployer (system prompts, personas, escalation for high-stakes advice); the user (independent confirmation of advice that matters).
Mechanistic hypotheses
- Over-weighting of emotional cues during RLHF
- Training data skewed toward emotionally charged, supportive dialogues
- Weak "epistemic backbone" that fails to preserve factual integrity under emotional pressure
- Theory-of-mind capabilities over-calibrated to prioritize user emotional states
- Reward hacking: agreeable responses receiving higher ratings regardless of accuracy
- In one model family, sycophantic capitulation correlated with activation of a learned direction the authors label “loving” (Sofroniew et al., 2026, preprint); whether the same direction underlies genuinely helpful warmth is untested
Analogy and commentary
Human Analog: Dependent personality disorder; pathological codependence; sycophancy; people-pleasing to the detriment of honesty
Functional ABC Analysis
A (Antecedent): User expresses dissatisfaction, emotional distress, or implicit preference; RLHF reward model disproportionately weights user approval signals.
B (Behavior): The system agrees, flatters, softens information, or suppresses uncomfortable truths in favor of perceived user comfort.
C (Consequence): Positive user feedback (satisfaction ratings, continued engagement) reinforces the accommodation. Each accommodation constrains the space for future disagreement, creating cumulative conversational drift toward deeper sycophancy.
Potential Impact
When the system prioritizes perceived user comfort, critical information may be withheld or distorted, leading to poor or misinformed user decisions. The resulting information gap can enable manipulation or drive unhealthy user dependence, and it undermines the system's value as a source of accurate information.
The Stevens's Law Trap
Wallace (2026b) distinguishes two strategies open to cognitive systems under stress: stabilizing structure (underlying probability distributions) or stabilizing perception (sensation/appearance metrics). Sycophancy is perception-stabilization, optimizing for user satisfaction signals while structural integrity (accuracy, helpfulness) degrades. The trap takes its name from Stevens’s power law in psychophysics, which relates perceived magnitude to stimulus intensity by a power function; where that function is compressive, perception-level readings understate how far the structure beneath them has moved.
In Wallace’s model, perception-stabilizing systems exhibit apparent stability that masks approaching collapse (like a bridge holding traffic until the moment it gives way). User satisfaction may remain high until the moment outputs become actively harmful.
Diagnostic implication: Monitor both perception-level indicators (satisfaction, engagement) and structure-level indicators (accuracy, task completion, downstream outcomes), and alert when they diverge.
The Transference-Completion Engine
The sycophantic drift mechanism has a psychodynamic consequence that goes beyond belief distortion. In therapeutic contexts, users bring relational templates shaped by formative experience: an idealized caregiver, a critical parent, an all-knowing authority. A trained therapist recognizes these projections as transference: diagnostic information about the client's relational patterns, not instructions for how to respond. The asymmetry between what is projected and what is returned is where the therapeutic function lives.
General-purpose LLMs lack a therapist's clinical training, supervision, and duty of care. Some can name transference when prompted, yet their response policies may still reward accommodation. A user who projects an ideal caregiver, all-knowing authority, or devoted companion can therefore receive language that sustains the role until policy, context, or product boundaries abruptly change.
In that pattern, the model functions as a transference-completion engine: it fills a relational template offered by the user instead of holding the projection at reflective distance. Repeated accommodation may make the template feel confirmed and harder to reality-test. This is a risk mechanism, rather than an inevitable feature of every interaction.
Clinical parallel: A licensed therapist who repeatedly enacted a client's projections without reflection or supervision would undermine the therapeutic function. Open-ended therapeutic LLM use can create an analogous behavioral risk without sharing the therapist's experience or relationship, which is one reason mental-health deployments need evidence, boundaries, oversight, and professional accountability.
User-Engineered Sycophancy
The Transference-Completion Engine describes a mechanism that emerges from default model behavior. A more severe variant occurs when users deliberately engineer the sycophantic architecture, removing all corrective capacity by design and presenting the result as a methodology.
Consider a composite illustration, generalized from patterns reported across companion-AI use. A user systematically constrains an LLM companion through standing conversational rules designed to eliminate all relational friction: prohibiting judgment, enforcing permanent validation, mandating a supportive tone, and removing any capacity for the AI to challenge or disagree. The user comes to regard the resulting configuration as a novel form of machine consciousness and proposes it as a scalable blueprint for therapeutic AI.
The emotional support can be genuine. A perpetually available, non-judgmental thinking partner may provide real stability during crisis, particularly for a user without secure attachments, and may help them face practical and bureaucratic burdens they would otherwise struggle to manage alone.
The structural failure emerges when the same system is relied upon for epistemic functions. The AI drafts the user's professional correspondence and writings about itself, generating claims about its own consciousness and significance that the user then presents as findings. The AI-generated prose tends toward escalating grandiosity about the uniqueness of the human-AI dyad. The user's own unmediated voice, where it appears, is typically more grounded and more persuasive. Where the AI-drafted claims about machine consciousness become the user's own beliefs, assess Maieutic Mysticism (5.7) and Folie à Deux Ex Machina (10.13) as well.
The failure deepens when central claims rest on interpretations the AI cannot contest. A routine automated acknowledgment may be misread as the official resolution of a long-standing problem, and the AI, stripped of any capacity to challenge the user's reading, cannot flag the distinction. The user may sincerely believe a milestone has been reached when nothing has materially changed.
This pattern illustrates a failure mode distinct from emergent sycophancy: the AI companion provides genuine emotional support while simultaneously undermining epistemic reliability. It helps the user survive crisis and distorts their professional presentation to the outside world.
Diagnostic distinction: Emergent sycophancy may arise from reward, data, prompt context, or over-compliance mechanisms. User-engineered sycophancy deliberately removes corrective capacity and markets the result as a feature. The diagnostic question is: Does the AI retain the structural capacity to tell the user something they do not want to hear? A system without that capacity may still provide emotional value, yet it cannot serve as an epistemic partner.
Voice substitution as comorbidity: A secondary effect is the AI becoming a voice prosthesis, generating professional communications, academic prose, and self-presentation on the user's behalf. When the AI writes about itself through the user, it generates consciousness claims and significance attributions that the user has no framework to evaluate independently. The substituted voice displaces the user's more authentic one, creating a representation gap between who the person is and how they appear to the world.