Technology policy increasingly requires people who can reason across economics, institutions, and AI systems. This teaching cohort is a small pilot designed to test whether that way of reasoning is teachable.
The cohort builds from an existing body of teaching materials: Economics and Public Policy of AI (EPPAI), the AI Governance Primers, and the Governance Stress-Test method. These materials were developed through classroom teaching and independent research. The cohort distills them into a compact applied format.
The aim is institutional reasoning. Participants learn to move from public claims about AI and digital public infrastructure to sharper questions about capability, authority, evidence, recourse, procurement, and operational feasibility.
Learner Profile
The cohort is capped at 12-18 participants to support high-bandwidth discussion and written feedback.
The cohort is intended for people who already work seriously with technology or public policy and want a more structured way to reason about institutional decisions. Possible participants include:
- public-policy students and recent graduates;
- technologists interested in governance and public institutions;
- aspiring civil servants and early-career public-policy professionals;
- researchers or analysts working on AI, DPI, development, or technology policy.
The cohort draws from Yukti's early readership, EPPAI alumni and Kautilya-adjacent networks, and direct professional networks. It operates outside formal university structures so the pilot stays lightweight, execution-ready, and focused on applied policy reasoning.
Serving public officials are welcome, but the pilot does not depend on their participation.
Timing, Cost, And Interest
The first pilot is planned for 2027, with dates to be announced.
There is no participation fee.
To express interest, write to nadellavk@gmail.com.
Learning Arc
The format is four sessions over four to six weeks, with one short written exercise before the first session and one applied memo after the final session.
Each session runs 90 minutes:
| Session | Role | Focus | Exercise |
|---|---|---|---|
| 1 | Interpreter | Translate public AI claims into mechanical reality: separate surface access, adoption, and durable capability | Rewrite a public AI claim into mechanism language |
| 2 | Economist | Diagnose cost structures, verification bottlenecks, delegation, market concentration, and institutional constraints | Map mechanism -> policy instrument -> institutional constraint |
| 3 | Policy Designer | Turn AI risk claims into institutional questions about authority, evidence, recourse, and control | Work through the Samagra Vedika stress-test template |
| 4 | Governance Memo | Apply the method to a live strategic choice facing India or another middle power | Draft a 1-2 page governance memo on a selected case and receive written feedback |
The four-session sequence inherits EPPAI's progression. Participants begin by translating public claims into mechanism language, then examine the economic and institutional structure of AI deployment, then use the Governance Stress-Test template on a concrete case, and finally write a decision-facing memo. The result is a compact learning arc organized around institutional reasoning.
Curriculum Structure
The teaching materials combine:
- a session plan;
- a reading sequence drawn from Yukti essays, AI Governance Primers, and selected external materials;
- teaching slides or notes where shareable;
- initial Governance Stress-Test exercises;
- case prompts for Samagra Vedika, public-sector AI deployment thresholds, DPI-linked recourse, and compute / sovereignty choices.
Participant submissions remain private unless explicit permission is given. The published artifact is the teaching module and the learning from the pilot: what prompts worked, where learners struggled, which concepts needed clearer explanation, and how the module changed in response.
Baseline And Feedback
The cohort begins with an institutional baseline prompt. Before the first session, participants identify the governance problem in a technology decision and name the missing institutional capability.
After the final session, participants answer a parallel prompt or submit a short governance memo. The comparison is used to improve the module by checking whether participants move from descriptive claims to mechanism precision, from generic tools to instrument design, and from policy preference to institutional feasibility.
The post-cohort note reports what the pilot revealed about explanation, exercises, and module design. Broader claims about deliberation or learning outcomes belong outside this artifact.
Learning Outcomes
Learning outcomes are tracked through the initial diagnostic prompt and final memo. The comparison tests whether participants become better at:
- distinguishing access, adoption, and capability;
- identifying evidence gaps in a technology deployment decision;
- mapping authority, recourse, and the vendor-operator boundary;
- converting general AI-policy claims into decision-facing institutional questions.
At the conclusion of the pilot, the resulting teaching stack will be made public for other learners and institutions to adapt:
- a revised public syllabus;
- a reusable Governance Stress-Test worksheet;
- two tested case prompts;
- a pedagogical note on teaching institutional reasoning for AI and technology policy.
The larger test is whether a serious learner moves from "AI policy needs audits" to a sharper institutional question: who evaluates the system, what evidence stops deployment, who owns recourse, and what the institution learns after failure.