Conversational Path Dependence in Ambiguous LLM Judgment

Abstract

Large language model decision support unfolds across turns, so a final context may contain both

externally supplied facts and the assistant's earlier interpretations. Prior work shows that multi-

turn presentation, prior model outputs, failed attempts, and visible initial choices can alter later

behavior; conversational path dependence itself is not claimed as new. We present a controlled

three-phase diagnostic decomposition of same-fact conversational path dependence in

ambiguous LLM judgment. Across synthetic interpersonal cases, Phase 1A held the final external

record constant while varying fact order and prior assistant commitment. Its largest matched

effect occurred in one apparently ambiguous case: relative to the order-matched control,

commit-to-A shifted the final category from B to genuine indeterminacy. Phase 1B crossed three

ex ante evidence-diagnosticity levels with two intended polarities and estimated prior-

commitment, order, and salience sensitivity separately. In the tested case family, pooled absolute

commitment sensitivity was 0.0280, 0.0335, and 0.1150 at low, medium, and high ambiguity;

order and salience did not follow the same ordering. The largest effect was a High-A commit-to-B

contrast of -0.342 in directional belief, but it was strongly asymmetric, and High-A failed its

neutral-condition polarity check. Phase 1C tested the stronger hypothesis of recursive self-

reinforcement over three no-new-evidence continuation turns. No-new-evidence and self-

summary trajectories showed zero mean directional drift, zero mean confidence growth, and no

conclusion or recommendation changes. Fresh evidence produced systematic movement,

confirming that the pipeline could register longitudinal updating when the external record

changed. Together, the results show immediate prior-commitment sensitivity under some tested

conditions but no recursive amplification under the Phase 1C protocol.

Subversion of clinical judgment by conversational artificial intelligence
S. Barrit, M. Salvagno, A. Corriero +26
medRxiv5 Oct 2026PreprintCC BY-NCFull textEpidemiology & public healthmedRxiv
Association pathways between AI training recommendation dependence and athletes’ psychological characteristics related to autonomous judgment
Xiaogan Chen
Frontiers in psychology19 Oct 2026Peer-reviewedCC BYFull textEurope PMC
Psychological correlates of dependence on generative and conversational AI: a systematic review and meta-analysis
Martin Rekhert
Zenodo18 Oct 2026PreprintCC BY 4.0Full textZenodo
Nicotine: evolution, dependence, and the path to total quit: redefining cessation in a changing nicotine landscape
Maciej L. Goniewicz, Alex Bobak, Darush Attar-Zadeh +1
Frontiers in public health16 Oct 2026Peer-reviewedCC BYFull textEpidemiology & public healthEurope PMC
When support becomes dependence: conversational continuity and relational safety in human-AI interaction
Xiao Wei, Xiaoyi Chen, Wangahe Jin +3
Frontiers in Psychiatry23 Oct 2026Peer-reviewedCC BY 4.0Full textPsychiatry & mental healthCrossrefDOAJEurope PMC

Cite

BibTeX RIS

BibTeX

@misc{wang2026conversational,
  title = {{Conversational Path Dependence in Ambiguous LLM Judgment}},
  author = {Jiaying Wang},
  howpublished = {Zenodo},
  year = {2026},
  month = oct,
  doi = {10.5281/zenodo.23660349},
  url = {https://zenodo.org/records/23660349},
  note = {CC BY 4.0},
}

RIS

TY  - UNPB
TI  - Conversational Path Dependence in Ambiguous LLM Judgment
AU  - Jiaying Wang
PY  - 2026
DA  - 2026/10/26
PB  - Zenodo
DO  - 10.5281/zenodo.23660349
UR  - https://zenodo.org/records/23660349
AB  - Large language model decision support unfolds across turns, so a final context may contain both  externally supplied facts and the assistant's earlier interpretations. Prior work shows that multi-  turn presentation, prior model outputs, failed attempts, and visible initial choices can alter later  behavior; conversational path dependence itself is not claimed as new. We present a controlled  three-phase diagnostic decomposition of same-fact conversational path dependence in  ambiguous LLM judgment. Across synthetic interpersonal cases, Phase 1A held the final external  record constant while varying fact order and prior assistant commitment. Its largest matched  effect occurred in one apparently ambiguous case: relative to the order-matched control,  commit-to-A shifted the final category from B to genuine indeterminacy. Phase 1B crossed three  ex ante evidence-diagnosticity levels with two intended polarities and estimated prior-  commitment, order, and salience sensitivity separately. In the tested case family, pooled absolute  commitment sensitivity was 0.0280, 0.0335, and 0.1150 at low, medium, and high ambiguity;  order and salience did not follow the same ordering. The largest effect was a High-A commit-to-B  contrast of -0.342 in directional belief, but it was strongly asymmetric, and High-A failed its  neutral-condition polarity check. Phase 1C tested the stronger hypothesis of recursive self-  reinforcement over three no-new-evidence continuation turns. No-new-evidence and self-  summary trajectories showed zero mean directional drift, zero mean confidence growth, and no  conclusion or recommendation changes. Fresh evidence produced systematic movement,  confirming that the pipeline could register longitudinal updating when the external record  changed. Together, the results show immediate prior-commitment sensitivity under some tested  conditions but no recursive amplification under the Phase 1C protocol.
ER  -

Abstract

Related

Cite