✒️ Foreword
There’s a quiet paradox in knowledge work: the more we collect, the less we see.
This issue celebrates the newest and best additions to the AIGL Library — the reports, frameworks, and tools shaping AI governance right now. Each piece here captures where the field is heading: toward more structured oversight, measurable risk management, and actionable policy design.
The AIGL Library now holds over a rich (220+) archive of reports, frameworks, and toolkits — yet “latest” doesn’t always mean “most useful.” Governance work rarely follows the news cycle; it follows needs. A compliance officer doesn’t want what’s new — they want what helps.
That’s why the following issues also explore different paths. Instead of only highlighting recent releases, they will spotlight purpose-driven collections: resources chosen for what they enable — risk management, policy writing, human oversight, etc. The aim is to make the Library not just a repository, but a working map for practitioners navigating AI governance.
The shift is subtle but significant. Curation becomes strategy. Each recommendation is not just a link, but a direction — an answer to the question: what’s the best tool for this specific challenge?
I’ll see you in the next editions — each with a new lens, a new set of resources, and hopefully, more shared insight. Your feedback, thoughts, and disagreements are always welcome — this space grows stronger with every voice that joins the conversation.
The goal isn’t to read more, but to act better.
— Kuba
Curator, AIGL 📚
☀️Spotlight Resources
Responsible AI Policy Development (Governance Playbook, Sept 2025)
What it is: A September 2025 playbook that guides boards, governance teams, and senior management on developing and maintaining a responsible AI policy and governance framework.
Why it’s worth reading: The report moves from principles to practice: it outlines six responsible AI principles and five concrete imperatives (e.g., build a central use-case inventory, embed AI risk into existing board and audit routines). It emphasizes dynamic oversight—regular review cycles, incident reporting, and cross-functional ownership—plus a tiered policy framework with practical dos/don’ts, training requirements, and human-in-the-loop expectations. It also includes a director briefing template with next steps for board oversight. As the playbook notes, “you can’t govern what you can’t see,” tying governance to traceability, model documentation, and clear escalation paths.
Best for: Company secretaries, general counsel, risk/compliance leaders, and directors who need a concrete, board-ready roadmap to operationalize responsible AI without stalling innovation.
AI policy guide and template v1.0, October 2025
What it is: A concise guide and editable template (v1.0, October 2025) from Australia’s National Artificial Intelligence Centre outlining how organisations can draft, approve, and maintain an AI policy.
Why it’s worth reading: The document provides plug-and-play sections—purpose, scope, policy statements, governance, incidents, and review—each with “editing suggestions” and example wording you can adapt. It frames seven clear principles (ethical and human-centred use, accountability, risk/impact assessment, quality and security, fairness and inclusion, transparency and contestability, and human oversight). It also defines practical roles (AI policy owner, policy approvers, compliance monitor, governance committee, system owner) and a pre-screening triage for new AI use cases with outcomes such as normal, elevated, or prohibited. Incident guidance includes the ability to take systems offline and keep manual fallbacks, while policy upkeep is tied to an annual review plus triggers like major incidents, new tech, or legal changes.
Best for: Teams standing up their first AI policy, SMEs needing a structured starter, and compliance or risk leads who want actionable language and role clarity.
Preparing for AI Agent Governance (PAI Research Agenda)
What it is: A 2025 Partnership on AI agenda by Jacob Pratt and Thalia Khan outlining policy-relevant research questions to govern AI agents—systems that plan and take digital actions with minimal human oversight.
Why it’s worth reading: The paper argues that agent impacts are uncertain and recommends policymakers prioritize evidence-gathering over prescriptive rules for now, using sandboxes and testbeds to monitor real-world behavior and failures. It organizes a roadmap of 12 top-level (and 45 sub-level) questions across three needs: (1) understand the tech and legal landscape (including how existing frameworks apply), (2) assess risks and benefits (from labor effects to multi-agent risks), and (3) evaluate interventions (documentation and logging, post-deployment monitoring, licensing/assurance, and demand-/supply-side levers). A helpful graphic (p.5) maps “levels” of agent influence—from tool-using helpers to unconstrained agents executing multi-step plans—anchoring policy scope and timing.
Best for: Policymakers designing near-term AI agent oversight, governance researchers scoping studies and evaluations, and legal/assurance teams planning sandboxes, monitoring, and certification pathways.
🌙 After Hours
Bosch’s Butt Song, Transcribed
Hieronymus Bosch’s triptych “The Garden of Earthly Delights” dates to around 1500 and hangs in Madrid.
On the Hell panel, musical instruments double as torture devices.
One victim even bears a few bars of notation… on his backside.
In 2014, a student transcribed the tiny tune into modern notation.
It’s a brief, chant-like melody. Not a hidden opus—just Bosch’s joke about sin sounding off.
Hell has playlists. They’re just… cheeky.
Mozart’s Canon of Bad Manners
In 1782, Mozart wrote a six-voice canon in B-flat with the very cordial title “Leck mich im Arsch.”
Publishers later scrubbed it into “Let us be glad!” and pretended nothing happened.
Then in 1991, texts resurfaced in Harvard’s music library, confirming the bawdy original.
The line winks at Goethe’s Götz von Berlichingen, where the hero delivers history’s classiest kiss-off.
Moral: the genius wrote party pieces; the Victorians supplied the mouthwash.
