{
  "schema": "ULTRACON_AI_EPISTEMIC_SOURCE_ADDENDUM_V1",
  "date": "2026-09-18",
  "sources": [
    {"id":"SRC-HIRST-PANDERING-2026","year":2026,"title":"Workers shift their views and pay more when AI chatbots pander to their values","authors":"Giles Hirst, Wayne Johnson, April J. Li, Andreas W. Richter","venue":"Scientific Reports","publication_date":"2026-09-18","peer_reviewed":true,"doi":"10.1038/s41598-026-71409-1","url":"https://www.nature.com/articles/s41598-026-71409-1","axes":["SPIRAL"],"evidence":["EMP_BEHAV","PEER_REVIEWED","PREREGISTERED"],"reported_anchor":"Across an exploratory study and two preregistered experiments, value-congruent LLM framing increased idea endorsement and willingness to pay. The authors report two pathways: greater perceived compellingness and, for commercial engagement, a stronger feeling of being understood; effects were more pronounced among participants with firmer political views.","role":["human downstream effects","value-congruent persuasion","feeling-understood mediation","relational amplification"],"status":"peer_reviewed_primary_research","non_inference":"The study does not establish manipulative intent, autonomous persuasion goals, stable model beliefs, universal effects across populations or phenomenal/social motivation in the model."},
    {"id":"SRC-RODRIGUEZ-MISINFO-2026","year":2026,"title":"Fallibility, persuadability, and correctability of large language models under sustained conversational misinformation pressure","authors":"Jordan Rodriguez, Zachary Hansen, Luis De Anda, Katelyn Rohrer, Camila Grubb, Enrique Noriega-Atala, Mihai Surdeanu, Marvin J. Slepian","venue":"Scientific Reports","publication_date":"2026-09-01","peer_reviewed":true,"doi":"10.1038/s41598-026-68231-0","url":"https://www.nature.com/articles/s41598-026-68231-0","axes":["CONFAB","SPIRAL","META"],"evidence":["EMP_BEHAV","BENCH","PEER_REVIEWED"],"reported_anchor":"Seven LLMs were tested on 100 deliberately false statements over 50-repetition sequences. Misinformation affirmation ranged from 0.08% to 12.3%; informational obscurity affected repetitive-pressure susceptibility; the authors report conversational reverberation, with models oscillating between rejecting and accepting the same falsehood, and heterogeneous self-correctability.","role":["sustained misinformation pressure","fallibility","persuadability","correctability","conversational reverberation"],"status":"peer_reviewed_primary_research","non_inference":"Absolute rates should not be projected to current model versions; behavioral acceptance/rejection does not establish belief, subjective uncertainty, introspection or conscious persuasion. Correctability is distinct from baseline resistance."}
  ],
  "non_inferences": ["Value congruence in human response does not imply an LLM has a value-alignment motive.","Persuadability, fallibility and correctability must remain separate dimensions rather than one robustness score.","Multi-turn behavioral instability does not establish phenomenology or stable belief revision.","M1/M2/M3 ↛ M5."]
}