Research map / Large language models / AI and society
Human-centered trust and agency
Model human trust, reliance, and agency when people work with AI systems and regulation.
161 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2018–2026) · an approach in AI and society, keyword cluster “human · systems · argue”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
38 papers in 2023–24 and 104 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 16.8% to 20.4%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2018 · Norm Conflict Resolution in Stochastic Domains (AAAI 2018 · 21 citations)
- 2019 · Updates in Human-AI Teams: Understanding and Addressing the Performance/Compatibility Tradeoff (AAAI 2019 · 283 citations)
- 2020 · Towards Socially Responsible AI: Cognitive Bias-Aware Multi-Objective Learning (AAAI 2020 · 14 citations)
- 2021 · Does Explainable Artificial Intelligence Improve Human Decision-Making? (AAAI 2021 · 106 citations)
- 2022 · Role of Human-AI Interaction in Selective Prediction (AAAI 2022 · 30 citations)
- 2023 · Modeling Human Trust and Reliance in AI-Assisted Decision Making: A Markovian Approach (AAAI 2023 · 18 citations)
- 2024 · Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties (AAAI 2024 · 31 citations)
- 2025 · Human-in-the-loop or AI-in-the-loop? Automate or Collaborate? (AAAI 2025 · 47 citations)
- 2026 · Formalising Human-in-the-Loop: Computational Reductions, Failure Modes, and Legal-Moral Responsibility (ICLR 2026 · 3 citations)
Earliest papers here
- Norm Conflict Resolution in Stochastic Domains (AAAI 2018 · 21 citations)
- How AI Wins Friends and Influences People in Repeated Games With Cheap Talk (AAAI 2018 · 11 citations)
- Updates in Human-AI Teams: Understanding and Addressing the Performance/Compatibility Tradeoff (AAAI 2019 · 283 citations)
What has been tried
52 papers here report 456 measured changes (a component added, removed or swapped, compared with a baseline): helped 57%, no clear effect 16%, hurt 23%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add relevance filtering to candidate generation | ValuePrism test situations · overall win rate vs GPT-4 (%) | 51.9 -> 55.5 | Helped | Value Kaleidoscope: Engaging AI with Pluralistic Human Values, Rights, and Duties (AAAI 2024 · 31 citations) |
| remove LoRA adapters | Audio classification average · average performance | 50.3 -> 38.3 | Hurt | Listen, Think, and Understand (ICLR 2024 · 18 citations) |
| Apply SmoothGrad to the original XAI methods | Faithfulness · faithfulness | No clear effect | Navigating the Maze of Explainable AI: A Systematic Approach to Evaluating Methods and Metrics (NeurIPS 2024 · 8 citations) | |
| Increase the specified log(Γ) toward the underlying confounding level | Synthetic data, homogeneous humans · policy regret | Helped | Confounding-Robust Deferral Policy Learning (AAAI 2025 · 0 citations) | |
| Add harm features to the aggregation | WildGuardTest vanilla, WildGuardTest adversarial, and WildJailbreak vanilla · F1 | 78.6 vs 90.9; 71.1 vs 79.6; 76.6 vs 88.8 | Helped | SafetyAnalyst: Interpretable, Transparent, and Steerable Safety Moderation for AI Behavior (ICML 2025 · 0 citations) |
| Impose the same regulatory penalty on the specialist as on the generalist | Regulation profiles (p, θ), outside abstention regions · qualitative equilibrium outcomes | No clear effect | Modeling the Economic Impacts of AI Openness Regulation (NeurIPS 2025 · 0 citations) | |
| Replace the probabilistic independent decision model with deterministic logistic regression | Delayed recommendation, 5 training instances · NLL | 0.430 -> 0.514 | Hurt | Decoding AI’s Nudge: A Unified Framework to Predict Human Behavior in AI-Assisted Decision Making (AAAI 2024 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 71 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Baselines
- Claude-3.5 Sonnet · 2 of 71 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 142 papers; from OpenAlex and the papers' affiliations).
- Stanford University · 13 papers
- Carnegie Mellon University · 7 papers
- University of California, Berkeley · 7 papers
- University of Washington · 6 papers
- Massachusetts Institute of Technology · 6 papers
- NVIDIA · 6 papers
- Google · 6 papers
- University of Oxford · 5 papers
- Google DeepMind · 5 papers
- Allen Institute for Artificial Intelligence · 4 papers
Use it
Download BibTeX of 34 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Social-science view of AI evaluation (159 papers)
- Data commons and human baselines (159 papers)
- Critiques of LLM claims and judging (104 papers)
- Benchmarking and reproducibility practice (104 papers)
- AI for health and public decisions (55 papers)
Elsewhere in the topic
- Teacher needs and AI curricula (45 papers)
- Classroom tools and student perceptions (40 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).