Skip to content
Apart Research
Apart Research
Contact us
Sprint projectFeb 2, 2026Manchester, United Kingdom

IG: A Unified Platform for Governed AI Agent Execution with Human-in-the-Loop Tool Verification

rng · Team RNG Inc

Submitted to The Technical AI Governance Challenge. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: IG: A Unified Platform for Governed AI Agent Execution with Human-in-the-Loop Tool Verification

Open in a new tab (opens in new tab)
Presentation (opens in new tab)
Share

Abstract: As AI systems transition from passive assistants to autonomous agents ca- pable of executing real-world actions, the governance challenge shifts from controlling model outputs to controlling model behaviors. This paper introduces Inference Gateway, a unified self-hosted platform that addresses critical gaps in AI governance through three novel contributions: a two-stage privacy-preserving pipeline, a multi-factor human-in- the-loop tool approval system with mobile device integration, and a provider-agnostic abstraction layer enabling consistent governance policies across heterogeneous AI infrastructure.

Link to Presentation

Reviews

Judging this Sprint?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

How much would this matter for AI safety if it worked? How innovative is it? For scores of 4-5: is this actually new to the field, or replicating recent work?

Scoring guide
  1. 1Negligible. No clear problem addressed, or no meaningful novelty.
  2. 2Limited. Addresses a real problem but with a generic or well-trodden approach. Incremental at best.
  3. 3Moderate. Clear problem with a reasonable approach; some novelty in framing or method beyond routine application of existing tools.
  4. 4Significant. Important problem with an original approach, or identifies a neglected problem area. A valuable contribution others could build on.
  5. 5Exceptional. Tackles a critical AI safety problem with a genuinely novel approach, or opens a new research direction. Clear theory of change. You'd be excited to share this with researchers in the area.

How sound are methodology, implementation, and findings?

Scoring guide
  1. 1Seriously flawed. Methodology broken, results uninterpretable, or implementation doesn't work.
  2. 2Weak. Approach has significant gaps: missing validation, flawed experimental design, or incomplete implementation.
  3. 3Competent. Technically solid given the short duration. Methodology makes sense, results are interpretable, limitations acknowledged, work builds toward clear conclusions.
  4. 4Strong. Thorough methodology with convincing validation. Results clearly support conclusions. Immediately useful for future work.
  5. 5Exceptional. Ambitious scope executed rigorously. Surprising findings, novel methods, or unusually robust validation.

How clearly are work, findings, and impact potential communicated?

Scoring guide
  1. 1Incomprehensible. Cannot determine what the project is actually claiming or doing.
  2. 2Hard to follow. Key information buried, missing, or diluted by excessive length. Significant effort to extract main points.
  3. 3Clear enough. Can understand the problem, approach, and results without undue effort. Core content clearly present: problem, method, findings, limitations.
  4. 4Well presented. Easy to follow, well-structured, appropriate level of detail. Target audience would get it quickly.
  5. 5Exceptionally clear. A pleasure to read. Complex ideas made accessible. Could serve as a model for how to present this type of work.

  1. There is a substantial body of work on training alignment and inference content filtering, but much less on controlling what agents actually do when they execute tools - which is what the paper covers. The design follows privileged access management, and treats high-stakes AI actions like financial transactions requiring multi-factor authentication. The two-stage privacy pipeline is also well designed and helps solve the inference privacy problem faced by many organisations. The risk-based approval escalation also greatly reduces the cost of human-in-the-loop procedures while retaining many of their advantages for reliability and security. This is clearly presented and well-engineered infrastructure that can help address an emerging operational need.

  2. Excellent framing of execution-time governance as a distinct and underserved layer! The analogy to privileged access management is compelling and well operationalized through the Duo integration and risk-based escalation matrix. To strengthen the work, I would want to see empirical evaluation, user testing, or latency benchmarks from the deployed system to demonstrate that the governance overhead is practical at scale.

Cite this project

@misc{rng2026ig,
  title = {{IG: A Unified Platform for Governed AI Agent Execution with Human-in-the-Loop Tool Verification}},
  author = {rng},
  year = {2026},
  month = feb,
  note = {Submitted to The Technical AI Governance Challenge, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/ig-a-unified-platform-for-governed-ai-agent-execution-with-humanintheloop-tool-verification-1ol1}},
  url = {https://apartresearch.com/sprints/projects/ig-a-unified-platform-for-governed-ai-agent-execution-with-humanintheloop-tool-verification-1ol1}
}

More from The Technical AI Governance Challenge

All 47 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026