Position: Trustworthy Model Context Protocol Enables Responsible Agentic AI!

Arjhun Swaminathan, Anika Hannemann
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:172208-172220, 2026.

Abstract

The Model Context Protocol (MCP) standardises AI agent-tool interaction, accelerating agentic AI adoption through interoperability. This presents an opportunity to embed trustworthiness. As a standard and an interface between agents and tools, MCP becomes a natural enforcement point; any improvements to it automatically propagate to all systems using it. Analyzing MCP through European Commission’s ethics guidelines for trustworthy AI, we identify three things: fundamental shifts in how trustworthiness works, critical challenges these shifts create, and strategic intervention points where protocol-level mechanisms can achieve ecosystem-wide impact. We argue that MCP’s architecture provides a foundation for trustworthiness and propose practical improvements to strengthen it. This position paper posits that building trustworthy MCP enables responsible agentic AI deployments.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-swaminathan26a, title = {Position: Trustworthy Model Context Protocol Enables Responsible Agentic {AI}!}, author = {Swaminathan, Arjhun and Hannemann, Anika}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {172208--172220}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/swaminathan26a/swaminathan26a.pdf}, url = {https://proceedings.mlr.press/v306/swaminathan26a.html}, abstract = {The Model Context Protocol (MCP) standardises AI agent-tool interaction, accelerating agentic AI adoption through interoperability. This presents an opportunity to embed trustworthiness. As a standard and an interface between agents and tools, MCP becomes a natural enforcement point; any improvements to it automatically propagate to all systems using it. Analyzing MCP through European Commission’s ethics guidelines for trustworthy AI, we identify three things: fundamental shifts in how trustworthiness works, critical challenges these shifts create, and strategic intervention points where protocol-level mechanisms can achieve ecosystem-wide impact. We argue that MCP’s architecture provides a foundation for trustworthiness and propose practical improvements to strengthen it. This position paper posits that building trustworthy MCP enables responsible agentic AI deployments.} }
Endnote
%0 Conference Paper %T Position: Trustworthy Model Context Protocol Enables Responsible Agentic AI! %A Arjhun Swaminathan %A Anika Hannemann %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-swaminathan26a %I PMLR %P 172208--172220 %U https://proceedings.mlr.press/v306/swaminathan26a.html %V 306 %X The Model Context Protocol (MCP) standardises AI agent-tool interaction, accelerating agentic AI adoption through interoperability. This presents an opportunity to embed trustworthiness. As a standard and an interface between agents and tools, MCP becomes a natural enforcement point; any improvements to it automatically propagate to all systems using it. Analyzing MCP through European Commission’s ethics guidelines for trustworthy AI, we identify three things: fundamental shifts in how trustworthiness works, critical challenges these shifts create, and strategic intervention points where protocol-level mechanisms can achieve ecosystem-wide impact. We argue that MCP’s architecture provides a foundation for trustworthiness and propose practical improvements to strengthen it. This position paper posits that building trustworthy MCP enables responsible agentic AI deployments.
APA
Swaminathan, A. & Hannemann, A.. (2026). Position: Trustworthy Model Context Protocol Enables Responsible Agentic AI!. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:172208-172220 Available from https://proceedings.mlr.press/v306/swaminathan26a.html.

Related Material