The 2025 AI Agent Index: Documenting Technical and Safety Features of Deployed Agentic AI Systems

Agentic AI systems are increasingly capable of performing professional and personal tasks with limited human involvement. However, tracking these developments is difficult because the AI agent ecosystem is complex, rapidly evolving, and inconsistently documented, posing obstacles to both researchers and policymakers. To address these challenges, this paper presents the 2025 AI Agent Index. The Index documents information regarding the origins, design, capabilities, ecosystem, and safety features of 30 state-of-the-art AI agents based on publicly available information and email correspondence with developers. In addition to documenting information about individual agents, the Index illuminates broader trends in the development of agents, their capabilities, and the level of transparency of developers. Notably, we find different transparency levels among agent developers and observe that most developers share little information about safety, evaluations, and societal impacts. The 2025 AI Agent Index is available online at https://aiagentindex.mit.edu

• Publication year

  2026

• Venue

  Unknown venue

• Publication date

  2026-02-19

• Fields of study

  Computer Science, Engineering, Environmental Science

• Identifiers
  arXiv 2602.17753
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Can't stop the hype: scrutinizing AI's realities

  2026cited by this paper
• Legal Alignment for Safe and Ethical AI

  2026cited by this paper
• An Economy of AI Agents

  2025cited by this paper
• From AI for Science to Agentic Science: A Survey on Autonomous Scientific Discovery

  2025cited by this paper
• AI Agents and the Law

  2025cited by this paper
• The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search

  2025cited by this paper
• Security Challenges in AI Agent Deployment: Insights from a Large Scale Public Competition

  2025cited by this paper
• OpenAgentSafety: A Comprehensive Framework for Evaluating Real-World AI Agent Safety

  2025cited by this paper
• Cybersecurity AI: The Dangerous Gap Between Automation and Autonomy

  2025cited by this paper
• The 2023 Foundation Model Transparency Index

  2025cited by this paper
• Technological Disruption in the Labor Market

  2025cited by this paper
• International AI Safety Report

  2025cited by this paper
• The AI Agent Index

  2025influential reference
• OCCULT: Evaluating Large Language Models for Offensive Cyber Operation Capabilities

  2025cited by this paper
• Agentic AI for Scientific Discovery: A Survey of Progress, Challenges, and Future Directions

  2025cited by this paper
• Impacts of generative artificial intelligence on the future of labor market: a systematic review

  2025cited by this paper
• A Survey of WebAgents: Towards Next-Generation AI Agents for Web Automation with Large Foundation Models

  2025cited by this paper
• How to evaluate control measures for LLM agents? A trajectory from today to superintelligence

  2025cited by this paper
• Levels of Autonomy for AI Agents

  2025influential reference
• Measurement to Meaning: A Validity-Centered Framework for AI Evaluation

  2025cited by this paper
• Position: Build Agent Advocates, Not Platform Agents

  2025cited by this paper
• Characterizing AI Agents for Alignment and Governance

  2025influential reference
• A Survey of AI Agent Protocols

  2025cited by this paper
• AI Behind Closed Doors: a Primer on The Governance of Internal Deployment

  2025cited by this paper
• The 2025 Foundation Model Transparency Index

  2025cited by this paper
• EvalCards: A Framework for Standardized Evaluation Reporting

  2025cited by this paper
• BrowseSafe: Understanding and Preventing Prompt Injection Within AI Browser Agents

  2025cited by this paper
• Remote Labor Index: Measuring AI Automation of Remote Work

  2025cited by this paper
• AI bots wrote and reviewed all papers at this conference

  2025cited by this paper
• GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks

  2025cited by this paper
• Holistic Agent Leaderboard: The Missing Infrastructure for AI Agent Evaluation

  2025cited by this paper
• A New Digital Divide? Coder Worldviews, the Slop Economy, and Democracy in the Age of AI

  2025cited by this paper
• A large-scale audit of dataset licensing and attribution in AI

  2024cited by this paper
• Visibility into AI Agents

  2024cited by this paper
• LLM Agents can Autonomously Hack Websites

  2024cited by this paper
• The artificial intelligence divide: Who is the most vulnerable?

  2024cited by this paper
• Safety Cases: How to Justify the Safety of Advanced AI Systems

  2024cited by this paper
• Evaluating Frontier Models for Dangerous Capabilities

  2024cited by this paper
• AI and the problem of knowledge collapse

  2024cited by this paper
• Governing AI Agents

  2024cited by this paper
• The Ethics of Advanced AI Assistants

  2024cited by this paper
• From "AI" to Probabilistic Automation: How Does Anthropomorphization of Technical Systems Descriptions Influence Trust?

  2024cited by this paper
• Systematic analysis of 32,111 AI model cards characterizes documentation practice in AI

  2024cited by this paper
• The 2024 Foundation Model Transparency Index

  2024cited by this paper
• Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?

  2024cited by this paper
• The AI Risk Repository: A Comprehensive Meta-Review, Database, and Taxonomy of Risks From Artificial Intelligence

  2024cited by this paper
• Improving Governance Outcomes Through AI Documentation: Bridging Theory and Practice

  2024cited by this paper
• Refusal-Trained LLMs Are Easily Jailbroken As Browser Agents

  2024cited by this paper
• Collapse or Thrive? Perils and Promises of Synthetic Data in a Self-Generating World

  2024cited by this paper
• Scalable watermarking for identifying large language model outputs

  2024cited by this paper
• Ecosystem Graphs: The Social Footprint of Foundation Models

  2023cited by this paper
• Identifying the Risks of LM Agents with an LM-Emulated Sandbox

  2023cited by this paper
• Harms from Increasingly Agentic Algorithmic Systems

  2023cited by this paper
• Accountability in an Algorithmic Society: Relationality, Responsibility, and Robustness in Machine Learning

  2022cited by this paper
• System Cards for AI-Based Decision-Making for Public Policy

  2022cited by this paper
• Future

  2021cited by this paper
• Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database

  2020cited by this paper
• From ethics washing to ethics bashing: a view on tech ethics from within moral philosophy

  2019cited by this paper
• FactSheets: Increasing trust in AI services through supplier's declarations of conformity

  2019cited by this paper
• Translating Principles into Practices of Digital Ethics: Five Risks of Being Unethical

  2019cited by this paper
• Datasheets for datasets

  2018cited by this paper
• Interpretable to Whom? A Role-based Model for Analyzing Interpretable Machine Learning Systems

  2018cited by this paper
• Model Cards for Model Reporting

  2018cited by this paper
• Index 5

  2009cited by this paper
• Modeling Rational Agents within a BDI-Architecture

  1997cited by this paper
• Agent-based software engineering

  1997cited by this paper
• Is it an Agent, or Just a Program?: A Taxonomy for Autonomous Agents

  1996cited by this paper
• Artificial life meets entertainment: lifelike autonomous agents

  1995cited by this paper
• Intelligent agents: theory and practice

  1995cited by this paper
• The Intentional Stance

  1989cited by this paper
• Membership

  1971cited by this paper
• An Introduction to Cybernetics

  1957cited by this paper
• Cybernetics, or Control and Communication in the Animal and the Machine.

  1949cited by this paper
• Behavior, Purpose and Teleology

  1943cited by this paper
• Generalized Disruption: Society, Work, and Property Rights in the Age of AI

  year unknowncited by this paper

• No citing papers are available for this paper.