Magentic-UI: Towards Human-in-the-loop Agentic Systems
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912510879203328 |
|---|---|
| author | Mozannar, Hussein Bansal, Gagan Tan, Cheng Fourney, Adam Dibia, Victor Chen, Jingya Gerrits, Jack Payne, Tyler Maldaner, Matheus Kunzler Grunde-McLaughlin, Madeleine Zhu, Eric Bassman, Griffin Alber, Jacob Chang, Peter Loynd, Ricky Niedtner, Friederike Kamar, Ece Murad, Maya Hosn, Rafah Amershi, Saleema |
| author_facet | Mozannar, Hussein Bansal, Gagan Tan, Cheng Fourney, Adam Dibia, Victor Chen, Jingya Gerrits, Jack Payne, Tyler Maldaner, Matheus Kunzler Grunde-McLaughlin, Madeleine Zhu, Eric Bassman, Griffin Alber, Jacob Chang, Peter Loynd, Ricky Niedtner, Friederike Kamar, Ece Murad, Maya Hosn, Rafah Amershi, Saleema |
| contents | AI agents powered by large language models are increasingly capable of autonomously completing complex, multi-step tasks using external tools. Yet, they still fall short of human-level performance in most domains including computer use, software development, and research. Their growing autonomy and ability to interact with the outside world, also introduces safety and security risks including potentially misaligned actions and adversarial manipulation. We argue that human-in-the-loop agentic systems offer a promising path forward, combining human oversight and control with AI efficiency to unlock productivity from imperfect systems. We introduce Magentic-UI, an open-source web interface for developing and studying human-agent interaction. Built on a flexible multi-agent architecture, Magentic-UI supports web browsing, code execution, and file manipulation, and can be extended with diverse tools via Model Context Protocol (MCP). Moreover, Magentic-UI presents six interaction mechanisms for enabling effective, low-cost human involvement: co-planning, co-tasking, multi-tasking, action guards, and long-term memory. We evaluate Magentic-UI across four dimensions: autonomous task completion on agentic benchmarks, simulated user testing of its interaction capabilities, qualitative studies with real users, and targeted safety assessments. Our findings highlight Magentic-UI's potential to advance safe and efficient human-agent collaboration. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_22358 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Magentic-UI: Towards Human-in-the-loop Agentic Systems Mozannar, Hussein Bansal, Gagan Tan, Cheng Fourney, Adam Dibia, Victor Chen, Jingya Gerrits, Jack Payne, Tyler Maldaner, Matheus Kunzler Grunde-McLaughlin, Madeleine Zhu, Eric Bassman, Griffin Alber, Jacob Chang, Peter Loynd, Ricky Niedtner, Friederike Kamar, Ece Murad, Maya Hosn, Rafah Amershi, Saleema Artificial Intelligence Human-Computer Interaction AI agents powered by large language models are increasingly capable of autonomously completing complex, multi-step tasks using external tools. Yet, they still fall short of human-level performance in most domains including computer use, software development, and research. Their growing autonomy and ability to interact with the outside world, also introduces safety and security risks including potentially misaligned actions and adversarial manipulation. We argue that human-in-the-loop agentic systems offer a promising path forward, combining human oversight and control with AI efficiency to unlock productivity from imperfect systems. We introduce Magentic-UI, an open-source web interface for developing and studying human-agent interaction. Built on a flexible multi-agent architecture, Magentic-UI supports web browsing, code execution, and file manipulation, and can be extended with diverse tools via Model Context Protocol (MCP). Moreover, Magentic-UI presents six interaction mechanisms for enabling effective, low-cost human involvement: co-planning, co-tasking, multi-tasking, action guards, and long-term memory. We evaluate Magentic-UI across four dimensions: autonomous task completion on agentic benchmarks, simulated user testing of its interaction capabilities, qualitative studies with real users, and targeted safety assessments. Our findings highlight Magentic-UI's potential to advance safe and efficient human-agent collaboration. |
| title | Magentic-UI: Towards Human-in-the-loop Agentic Systems |
| topic | Artificial Intelligence Human-Computer Interaction |
| url | https://arxiv.org/abs/2507.22358 |