Social sciences · OECD FoS 5.4 · China MoE 0303 社会学 / 0302 政治学
Sociology & political science
社会学与政治学
L1Assistive → Steady Score 1.30 · Last reviewed 2026-10-22 · Next review 2027-01-19
BIGGEST BREAKTHROUGH 2023-02-21 AI · majorField · major
“Out of One, Many”: LMs simulate human samples (Political Analysis) ↗
Introduced LLM “silicon samples”
Why this level
“Silicon samples” (LLM-simulated respondents, Political Analysis 2023) spark debate but reliability is doubtful (Nat. Hum. Behav. 2026); main uses are text coding and survey support.
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
Composite 0.75–1.49: AI is an aid in parts of research or practice, with cases or surveys, but no recognised AI-key result. All levels →
Research directions
Silicon samples & simulationLarge-scale text codingPublic opinion
Papers
- Simulated responses from large language models are weak proxies for human economic preferences across 74 countries
74-country study: silicon samples are weak proxies
2026-10-12 · Nature Human Behaviour
- Towards scalable AI-assisted pre-bunking of election misinformation: evidence from a pre-registered US panel experiment
AI: once experts design the prompt, LLMs mass-produce pre-bunks without per-article review. Field: pre-registered evidence for scalable AI counter-misinformation
2026-08-05 · Royal Society Open Science
- The levers of political persuasion with conversational artificial intelligence
Largest AI persuasion experiment to date
2025-12-04 · Science
- Can Generative AI improve social science?
Sociologist’s systematic assessment
2024-05-09 · Proceedings of the National Academy of Sciences
- Can Large Language Models Transform Computational Social
Science?
Benchmarks LLMs on 25 social-science coding tasks
2024-03-01 · Computational Linguistics
- Out of One, Many: Using Language Models to Simulate Human Samples
Introduced LLM “silicon samples”
2023-02-21 · Political Analysis
- Measuring the predictability of life outcomes with a scientific mass collaboration
Key negative result: ML barely predicts life outcomes
2020-03-30 · Proceedings of the National Academy of Sciences
- Where Do LLMs Learn to Recommend AI?
Finds LLMs often recommend AI-related options and that instruction tuning sharply increases this tendency.
2026-10-21 · arXiv · Benaya Trabelsi, Jonathan Shaki, Sarit Kraus
- Situating Vulnerability: An Ethical Governance Framework for Anthropomorphic AI Interaction
An ethical governance framework for anthropomorphic AI that situates user vulnerability rather than applying uniform rules.
2026-10-21 · arXiv · Hanhui Xu, Jiacheng Ji, Yuhan Wang, Haoan Jin
- When LLMs Manipulate: Measuring Behavioral, Social, and Temporal Effects
A 120-participant study finds manipulative LLM framing shifts trust and compliance, with effects propagating socially.
2026-10-20 · arXiv · Sadia Asif, Israel Antonio Rosales Laguan, Shumaila Asif, Haris Khan
- EpiLens: Lightweight RAG for Epidemiological Knowledge Acquisition
Open-source RAG package for mapping epidemiological literature, evaluated on 2,297 papers against 214 labels.
2026-10-19 · arXiv · Vincenzo Perri, Ciro Cattuto, Daniela Paolotti
- Where Does Rhetorical Distortion Live? A Multi-Platform, Multi-Language Measurement of Influencer Content
Measures five dimensions of rhetorical distortion in influencer content across platforms and languages.
2026-10-19 · arXiv · Xucheng Yu, Jiangping Chen, Haohan Wang
- Aligning Language Models with Indigenous Community Cosmovision through Constitutional AI
Builds a 12-principle cultural constitution from interviews with Indigenous leaders to align LLMs.
2026-10-18 · arXiv · Silvia Fernández-Sabido, Jorge Reyes-Magaña, Jesús Adrián Chay Tun, Jean Carlos Buenfil
- Do LLMs Really Lack Difference Awareness? Uncovering Heuristic Substitution in Group-Differentiating Fairness
Argues LLMs substitute an equal-treatment heuristic for difference awareness and proposes a two-stage fix.
2026-10-18 · arXiv · Jingshen Zhang, Bo Wang, Yanlin Fu, Yuewen Dang
- Scam Town: Can Scam Education Spread Like a Vaccine?
An LLM-simulated community tests whether word-of-mouth scam warnings spread protection like a vaccine.
2026-10-18 · arXiv · Simiao Ren, Tommy Duong, Xingyu Shen, Jiaqi Gan
- From arXiv, bioRxiv, medRxiv and OpenAlex (last ~30 days); summaries written by AIMap from the abstract.
Milestones 7
Big glowing nodes are leaps: AI leap cool · Field leap warm · double ring = both
-
Paper
74-country study: LLM simulated responses are weak proxies for humans (Nat. Hum. Behav.)74-country study: silicon samples are weak proxies
Nature Human Behaviour -
Paper
Pre-registered US panel experiment: AI-generated pre-bunking reduces belief in five election rumours, effects persist a week (R. Soc. Open Sci.)AI: once experts design the prompt, LLMs mass-produce pre-bunks without per-article review. Field: pre-registered evidence for scalable AI counter-misinformation
Royal Society Open Science -
Paper
Hackenburg et al.: The levers of political persuasion with conversational AI (Science)Largest AI persuasion experiment to date
Science -
Paper
Bail: Can generative AI improve social science? (PNAS)Sociologist’s systematic assessment
PNAS -
Paper
Can LLMs transform computational social science? (Computational Linguistics)Benchmarks LLMs on 25 social-science coding tasks
Computational Linguistics -
Paper
“Out of One, Many”: LMs simulate human samples (Political Analysis)Introduced LLM “silicon samples”
Political Analysis -
Paper
Salganik et al.: Measuring the predictability of life outcomes with a scientific mass collaboration (PNAS)Key negative result: ML barely predicts life outcomes
PNAS
No milestones match this filter.
More papers & repos 6
| Type | Title | Date |
|---|---|---|
| Paper | Simulated responses from large language models are weak proxies for human economic preferences across 74 countries | 2026-10-12 |
| Paper | The levers of political persuasion with conversational artificial intelligence | 2025-12-04 |
| Paper | Can Generative AI improve social science? | 2024-05-09 |
| Paper | Can Large Language Models Transform Computational Social Science? | 2024-03-01 |
| Paper | Out of One, Many: Using Language Models to Simulate Human Samples | 2023-02-21 |
| Paper | Measuring the predictability of life outcomes with a scientific mass collaboration | 2020-03-30 |
Latest news 5
- How governments worldwide are shaping what AI can say about politicsThe Conversationen
- Democrats prepare to battle over AI regulationTransformeren
- Democrats want to lead on AI – but can’t agree howTransformeren