Social sciences · OECD FoS 5.1 · China MoE 0402 心理学
Psychology
心理学
L2Local breakthroughs → Steady Score 1.65 · Last reviewed 2026-10-21 · Next review 2027-01-19
BIGGEST BREAKTHROUGH 2025-07-02 AI · majorField · major
Centaur: a foundation model of human cognition (Nature) ↗
Cognitive foundation model; disputed
Why this level
Centaur, fine-tuned on 10M+ human choices, predicts behaviour better than classic cognitive models (Nature 2025), though evaluations show divergence from humans; AI dialogues durably reduced conspiracy beliefs (Science 2024).
S = 0.30·method + 0.35·discovery + 0.25·adoption + 0.10·education (education scores 0 when there is no evidence; if empty, the other weights are renormalised)
Sources
- nature.com/articles/s41586-025-09215-4
- arxiv.org/abs/2508.07887
- science.org/doi/10.1126/science.adq1814
Composite 1.5–2.49: at least one peer-reviewed/officially verified AI-key result, or one dimension at 3; routine workflow largely unchanged. All levels →
Research directions
Cognitive foundation modelsAI therapy evaluationLLM participantsAI persuasion interventions
Papers
- AI-generated vs. human therapist feedback in internet-based CBT for social anxiety disorder: A six-arm randomized controlled trial with deception design
AI: psychologist-reviewed AI feedback showed no efficacy disadvantage; Field: non-inferiority not established
2026-10-16 · Internet Interventions
- A reporting checklist for large language models in behavioural science
AI: standardises reporting of model versions, prompts and settings; Field: first consensus reporting standard for LLM research in behavioural science
2026-06-09 · Nature Human Behaviour
- Efficacy of a Conversational AI Agent for Psychiatric Symptoms and Digital Therapeutic Alliance
Large RCT; long-term follow-up needed
2026-04-14 · JAMA Network Open
- A cognitive layer architecture to support large-language model performance in psychotherapy interactions
Double-blind, real-world-scale evaluation
2026-03-12 · Nature Medicine
- A foundation model to predict and capture human cognition
Cognitive foundation model; disputed
2025-07-02 · Nature
- Randomized Trial of a Generative AI Chatbot for Mental Health Treatment
First RCT of a fully generative therapy chatbot
2025-03-27 · NEJM AI
- Randomized Trial of a Generative AI Chatbot for Mental Health Treatment
First RCT of a generative-AI therapy chatbot
2025-03-27 · NEJM AI
- Durably reducing conspiracy beliefs through dialogues with AI
~20% belief reduction lasting two months
2024-09-13 · Science
- Using cognitive psychology to understand GPT-3
Opened the “LLM as participant” line
2023-02-02 · Proceedings of the National Academy of Sciences
- Same Probabilities, Different Advice: Gain-Loss Framing Effects in High-Stakes LLM Decision Support
FrameBench-HPS tests whether LLM high-stakes advice shifts when the same probability is framed as gain or loss.
2026-10-21 · arXiv · Yuyao Zhang, Yakun Cui, Han Zhu, Jinsong Lu
- Scaling Therapy Supervision: Assessing Psychotherapy Fidelity Using Large Language Models
Uses LLMs to automate psychotherapy fidelity assessment and generate actionable feedback for clinicians.
2026-10-20 · arXiv · Arvind Pillai, Alexis Roach, Jason Holden, Varsha D Badal
- Reading Too Much Into It? Interpretive Preferences and Commitment in LLM Mentalisation under Ambiguity
MentalLoom shows user distress disclosures and repeated questioning shift LLM interpretations and certainty without new evidence.
2026-10-20 · arXiv · Stephanie Fong, Yiwen Jiang, Qingyang Xu, Zimu Wang
- BAIT: Can LLMs Find What Biases a Reader? A Benchmark for Cognitive Bias Triggers in News
A benchmark of cognitive-bias triggers in news; LLMs approach human agreement on presence but miss spans.
2026-10-20 · arXiv · Yousra Fettach, Tijl De Bie
- Why Do LLMs Accept Generics? Explaining Acceptability using Psycholinguistic and Model-Derived Factors
LLMs weigh prevalence and distinctiveness like humans but rely most on factual truth when judging generics.
2026-10-20 · arXiv · Mervat Abassy, Zhaoxin Feng, Tatsuki Kuribayashi, Kentaro Inui
- Cognitive Map in Large Language Models
Models above ~10B parameters infer unstated relations and form abstract, cognitive-map-like relational representations.
2026-10-19 · arXiv · Albert Qin, Jonathan Cohen, Gautam Reddy
- Switching intention toward conversational AI for health information seeking: a push–pull–mooring analysis
Survey of 2,876 users explains switching from social media to conversational AI for health information.
2026-10-19 · OpenAlex · Frontiers in Psychology · Zhaoqi Zhou, Yan Lin, Quanzhi Chen, Yanting Wu
- Reporting crime to police via human and AI live chats: the salience of procedural justice and outcome favourability
An online experiment (N=648) compares perceived fairness of crime reporting via human vs AI live chat.
2026-10-18 · OpenAlex · Police Practice and Research · Ben Bradford, Arabella Kyprianides, Helen Wells, Elizabeth Aston
- From arXiv, bioRxiv, medRxiv and OpenAlex (last ~30 days); summaries written by AIMap from the abstract.
Milestones 9
Big glowing nodes are leaps: AI leap cool · Field leap warm · double ring = both
-
Clinical
Six-arm RCT: AI-generated vs therapist feedback in internet CBT for social anxiety (Internet Interventions)AI: psychologist-reviewed AI feedback showed no efficacy disadvantage; Field: non-inferiority not established
Internet Interventions -
Policy
A reporting checklist for large language models in behavioural science (Nat. Hum. Behav.)AI: standardises reporting of model versions, prompts and settings; Field: first consensus reporting standard for LLM research in behavioural science
Nature Human Behaviour -
Clinical
JAMA Netw Open RCT (n=995): conversational AI agent improves psychiatric symptoms vs group therapy and waitlistLarge RCT; long-term follow-up needed
JAMA Network Open -
Clinical
Cognitive-layer architecture for LLM psychotherapy: 227-participant double-blind evaluation (Nature Medicine)Double-blind, real-world-scale evaluation
Nature Medicine -
Paper
Centaur: a foundation model of human cognition (Nature)Cognitive foundation model; disputed
NatureContested -
Clinical
Therabot RCT for depression, anxiety and eating-disorder symptoms (NEJM AI)First RCT of a fully generative therapy chatbot
NEJM AI -
Paper
Therabot RCT: generative AI chatbot for mental health treatment (NEJM AI)First RCT of a generative-AI therapy chatbot
NEJM AI -
Paper
Dialogues with AI durably reduce conspiracy beliefs (Science)~20% belief reduction lasting two months
Science -
Paper
Using cognitive psychology to understand GPT-3 (PNAS)Opened the “LLM as participant” line
PNAS
No milestones match this filter.
More papers & repos 8
| Type | Title | Date |
|---|---|---|
| Paper | A reporting checklist for large language models in behavioural science | 2026-06-09 |
| Paper | Efficacy of a Conversational AI Agent for Psychiatric Symptoms and Digital Therapeutic Alliance | 2026-04-14 |
| Paper | A cognitive layer architecture to support large-language model performance in psychotherapy interactions | 2026-03-12 |
| Paper | A foundation model to predict and capture human cognition | 2025-07-02 |
| Paper | Randomized Trial of a Generative AI Chatbot for Mental Health Treatment | 2025-03-27 |
| Paper | Randomized Trial of a Generative AI Chatbot for Mental Health Treatment | 2025-03-27 |
| Paper | Durably reducing conspiracy beliefs through dialogues with AI | 2024-09-13 |
| Paper | Using cognitive psychology to understand GPT-3 | 2023-02-02 |