BAI Europe · The Series Behind Winter School 2027

Beyond Automation:
The Human Stake in AI

Ten articles on identity, trust, and governance in agentic systems — built around the live cases anchoring BAI Winter School 2027 in Cape Town: MyMzansi, SARS Modernisation 3.0, and the EU Digital Identity Wallet.

Publishing 1st & 3rd Wednesday · 19 Aug 2026 – 6 Jan 2027 Opens with The Missing Identity
Subscribe to the Series See the Ten Articles

Identity, trust, and governance — before the cohort arrives

Ten articles, each standing alone, building the case that Winter School 2027 makes over ten residential days: agentic systems raise questions about identity, trust infrastructure, and accountability that no compliance checklist answers on its own. The series opens with The Missing Identity and closes on Trust by Design — the programme's own capstone workshop — timed to close out well ahead of the 2027 cohort's departure for Cape Town.

Who reads it: graduate students and working professionals considering Winter School 2027, technology and innovation leaders governing agentic deployments today, and anyone deciding whether MyMzansi, SARS Modernisation 3.0, or the EU Digital Identity Wallet has something to teach their own organization.

91% vs 10%

Of organizations already use AI agents in production, but only one in ten has a strategy for managing the non-human identities those agents create. (Okta, 2026)

97% / 6%

Of security leaders expect a material AI-agent incident this year — while only 6% of security budgets are allocated to the risk. (Lasso Security, 2026)

Six reasons this is an HP (Human Potential) newsletter, not an AI adoption one

An Identity You Can Grow From

Agents already outnumber human identities 25–50x in many enterprises. Naming what's automated is the first step to naming what's still worth developing in yourself.

Judgment, Practiced Not Outsourced

Zero Trust verifies who acted. It doesn’t build the judgment a human still needs to exercise — and like any other capability, that one only grows with deliberate practice.

Real Cases to Learn From, Not Theorize About

MyMzansi, SARS Modernisation 3.0, and the EU Digital Identity Wallet are live and running now — real institutions building real capability, not a hypothetical for a classroom.

Agency as a Capability, Not Just a Setting

The EUDI Wallet's selective disclosure shows personal agency can be designed for — a working model for the same discipline a person has to build for themselves.

A Ladder, Not Just a Checklist

Maker, Checker, Accountable Principal isn’t only an org chart — it’s a growth path. The series maps what it actually takes to earn your way to the third rung.

Evidence Over Hype

Every claim traces to a dated, named 2026 source. Developing real judgment about agentic AI starts with evidence — not the next adoption headline.

Identity, Trust, Governance

Act I · Identity Who is acting — and whether the theory behind human oversight even applies
01

The Missing Identity

91% of organizations use AI agents; only 10% have a strategy for managing the non-human identities those agents create. The series opens on the question no one has answered yet.

19 Aug 2026Publishing
02

Zero Trust for Systems That Act

The Cloud Security Alliance's Agentic Trust Framework and NIST SP 800-207 extend “never trust, always verify” to non-human principals.

2 Sep 2026Coming
03

The Principal Who Doesn’t Know They’re the Principal

Principal-agent theory assumes agents respond to incentives and sanctions. 2026 scholarship argues that assumption breaks down entirely for AI agents.

16 Sep 2026Coming
Act II · Trust Three live cases, each asking whether a human retained a real choice
04

Cape Town’s Digital Front Door

MyMzansi built a national identity front door in ten weeks — deliberately voluntary. Does an opt-out clause preserve real agency once a system becomes “the only gateway”?

7 Oct 2026Coming
05

Tax Just Happens

SARS Modernisation 3.0 pairs biometric identity with agentic AI that autonomously flags, assesses, and collects — against a record R2.01 trillion collected in 2025/26.

21 Oct 2026Coming
06

Twenty-Seven Wallets, One Continent

The EUDI Wallet's selective-disclosure design builds agency into the architecture itself — but EU pilots already flag the “cookie problem” of consent fatigue.

4 Nov 2026Coming
07

Who Still Holds the Veto

97% of security leaders expect a material AI-agent incident this year; only 6% of budgets are allocated to it, and 68% can't tell human activity from agent activity in their own logs.

18 Nov 2026Coming
Act III · Governance Who answers for it — and the architecture that makes the answer real
08

Which Half of the Job Survives

Which human judgments an organization deliberately keeps — judgment under ambiguity, institutional memory, relational trust — versus which it lets erode by default.

2 Dec 2026Coming
09

The Maker, the Checker, and the Accountable Principal

BAI's own accountability architecture, made explicit — the structural answer to every veto-in-name-only case built across Acts I and II.

16 Dec 2026Coming
10

Trust by Design

Capstone, named for the programme's own Day 9 workshop. Synthesizes identity, trust, and governance into a portable decision tool — landing three weeks ahead of the 2027 cohort's departure for Cape Town.

6 Jan 2027Coming

This series is the reading list for Winter School 2027

Ten days in Cape Town, February 2027 — MyMzansi, SARS Modernisation 3.0, and the EUDI Wallet as live case studies, taught by the people building them. Application deadline: December 2026.

Visit baiwinter.com