Research map / Large language models
AI and society: research map
913 papers on AI and society in Large language models, from ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR (2016–2026), grouped into 2 clusters and 8 approaches. The busiest year so far is 2025.
What it is
How AI systems, especially language models, affect people and institutions: education, trust and human agency, how AI evaluation is done and reported, and AI in health and public decisions. Papers here often come from the social sciences and critique common practice.
Terms you will see on this page:
- Human-AI interaction
- how people use, trust and depend on AI systems.
- Benchmark validity
- whether a benchmark actually measures the ability it claims to.
- Reproducibility
- whether others can repeat a result with the same data and code.
Start reading with the most cited papers here:
- SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems (NeurIPS 2019 · 984 citations)
- CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generation (NeurIPS 2021 · 408 citations)
Within Large language models, its share shrank from 11.4% in 2023–24 to 7.4% in 2025–26 (226 → 511 papers at ICML, NeurIPS, ICLR, CVPR and AAAI, the venues with data for all four years).
Explore AI and society in the interactive map
Working on something in this topic? Describe your idea in scime atlas to see which approach it falls under, the closest papers by meaning and how crowded the spot has become.
What changed recently
Approaches whose share of the topic moved most between 2023–24 and 2025–26 (ICML, NeurIPS, ICLR, CVPR and AAAI):
- Data commons and human baselines rising: 13% → 20% of the topic (29 → 104 papers)
- Critiques of LLM claims and judging rising: 11% → 15% of the topic (24 → 79 papers)
- Social-science view of AI evaluation rising: 17% → 21% of the topic (38 → 109 papers)
- Benchmarking and reproducibility practice falling: 17% → 9% of the topic (38 → 47 papers)
- Teacher needs and AI curricula falling: 10% → 4% of the topic (23 → 22 papers)
Papers from 2025–2026 that are already cited most:
- Human-in-the-loop or AI-in-the-loop? Automate or Collaborate? (AAAI 2025 · 47 citations)
- Why AI Is WEIRD and Shouldn't Be This Way: Towards AI for Everyone, with Everyone, by Everyone (AAAI 2025 · 22 citations)
- AI-Driven Virtual Teacher for Enhanced Educational Efficiency: Leveraging Large Pretrain Models for Autonomous Error Analysis and Correction (AAAI 2025 · 9 citations)
- SNAP: Semantic Stories for Next Activity Prediction (AAAI 2025 · 6 citations)
Approaches and key papers
AI, people and evaluation practice · 826 papers
Keywords: human, systems, argue
Approaches in this cluster:
- Human-centered trust and agency (166 papers)
Model human trust, reliance, and agency when people work with AI systems and regulation. - Social-science view of AI evaluation (170 papers)
Argue for rigorous, socially grounded measurement and broader evaluation of AI systems. - Data commons and human baselines (167 papers)
Measure AI's societal footprint, data consent, and rigor of human baselines. - Critiques of LLM claims and judging (104 papers)
Examine validity of LLM research claims, LLM-as-judge use, and research benchmarks. - Benchmarking and reproducibility practice (139 papers)
Promote common task frameworks, dataset documentation, and replication for scientific ML. - AI for health and public decisions (80 papers)
Apply AI frameworks to health monitoring, mental health, and equitable public decisions.
Most cited and most cited since 2024:
- SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems (NeurIPS 2019 · 984 citations)
- CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generation (NeurIPS 2021 · 408 citations)
- MemoryBank: Enhancing Large Language Models with Long-Term Memory (AAAI 2024 · 201 citations)
- Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews (ICML 2024 · 70 citations)
AI in education · 87 papers
Keywords: students, education, teaching
Approaches in this cluster:
- Teacher needs and AI curricula (46 papers)
Survey and design AI literacy education for K-12 teachers and educator preparation. - Classroom tools and student perceptions (41 papers)
Build and evaluate interactive classroom tools and study how students use generative AI.
Most cited and most cited since 2024:
- Students’ Perceptions and Preferences of Generative Artificial Intelligence Feedback for Programming (AAAI 2024 · 46 citations)
- Co-designing AI Education Curriculum with Cross-Disciplinary High School Teachers (AAAI 2024 · 37 citations)
- An Effectiveness Study of Teacher-Led AI Literacy Curriculum in K-12 Classrooms (AAAI 2024 · 29 citations)
- Build Your Own Robot Friend: An Open-Source Learning Module for Accessible and Engaging AI Education (AAAI 2024 · 20 citations)
Related topics in Large language models
- Reasoning and chain of thought (1,574)
- Instruction and fine-tuning (888)
- Retrieval and knowledge (1,084)
- Agents, code and math (1,748)
- Alignment and preferences (672)
- Language, safety and interpretability (3,002)
