A working paper, published in the open. Criticism, corrections and clarifications are welcome.
Send yours → Library /Everything Library · 90 resources
Everything on the shelf Every entry says what it gives you, what it leaves out, how long it takes, and which parts of the framework it speaks to, in red. Filter by any of them, or start with one of the four paths .
Showing 90 of 90
What you’ll get How a large emerging economy governs AI mainly through existing laws: seven 'sutras', six pillars, and new coordinating institutions.
What it leaves out Government summary of the guidelines, not the full text; India-specific. Speaks to: Institutions, Infrastructures, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
What you’ll get The international scientific consensus on what general-purpose AI can do and its risks: malicious use, malfunctions and systemic risks.
What it leaves out Deliberately makes no specific policy recommendations. Speaks to: Safeguards, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · International
about 30 min (outline) Intermediate Free
What you’ll get Japan's soft-law model: ten common principles plus specific duties for AI developers, providers and business users.
What it leaves out Voluntary guidance; outline only, some appendix sections Japanese-only. Speaks to: Safeguards, Incentives, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
What you’ll get Asia's first comprehensive AI statute: duties for high-impact AI, generative-AI labelling, a domestic-representative rule and modest fines.
What it leaves out Many details left to Presidential Decree; Korea-specific; unofficial translation. Speaks to: Safeguards, Institutions, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
about 15 min Intermediate Free
What you’ll get The founding text for the UN's two new AI bodies: a 40-member scientific panel and a global dialogue on AI governance.
What it leaves out Procedural terms of reference; sets up bodies but creates no rules for AI itself. Speaks to: Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
about 5 hrs Intermediate Free
What you’ll get Africa's human-rights body on AI risks for the continent: surveillance, bias, dependence on foreign systems, data exploitation and regulatory gaps.
What it leaves out Long study; recommendations section not read for this entry. Speaks to: Power, Safeguards, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
about 10 min (Commission overview) Beginner Free
What you’ll get The clearest official entry point to the EU AI Act: its four risk tiers, general-purpose AI rules, enforcement bodies and dates.
What it leaves out Applies to the EU market; the overview simplifies a long legal text. Speaks to: Safeguards, Institutions, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
about 10 min (brochure) Beginner Free
What you’ll get A plain summary of the first binding international AI treaty: principles, remedies for affected people, and risk and impact assessment duties.
What it leaves out Brochure, not the treaty text; national security and defence largely outside scope. Speaks to: Safeguards, Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
about 30 min (executive summary); 4-5 hrs full Beginner Free
What you’ll get The UN's case for global AI governance and seven concrete proposals, including a scientific panel, policy dialogue, capacity network and global fund.
What it leaves out Proposals, not binding rules; later UN decisions implemented only some of them. Speaks to: Institutions, Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
about 2 hrs Intermediate Free
What you’ll get The main example of a sector-led model: five principles applied by existing regulators rather than a new AI law or regulator.
What it leaves out UK-specific 2023 policy proposal; later UK policy developments not covered. Speaks to: Institutions, Incentives, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
What you’ll get A practical, widely used structure (Govern, Map, Measure, Manage) for organisations to identify and manage AI risks.
What it leaves out Voluntary and US-origin; process guidance rather than legal requirements. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
What you’ll get China's stated position on global AI governance, stressing sovereignty, equal development rights, and more voice for developing countries.
What it leaves out Short statement of positions; no implementation details or domestic rules. Speaks to: Power, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
What you’ll get The G7's voluntary list of actions for developers of advanced AI, applied across design, deployment and use on a risk basis.
What it leaves out Voluntary and aimed at developers of advanced systems; G7-led. Speaks to: Safeguards, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
about 2 hrs (paid) Advanced Paid
What you’ll get The certifiable international standard for running an AI management system inside an organisation.
What it leaves out Paid (CHF 225); management-system requirements, not rules about specific AI uses. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
What you’ll get Latin America's leading attempt at a comprehensive AI law, with risk classification, liability and public governance provisions.
What it leaves out Still a bill; Portuguese only; final content depends on the Chamber of Deputies. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
What you’ll get A short, readable statement where the US, China, the EU, India, Kenya, Nigeria and others jointly named frontier AI risks.
What it leaves out Political declaration with no binding commitments or follow-up mechanism in the text. Speaks to: Safeguards, Power, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
What you’ll get A short brochure summarising the global ethics standard: its values, principles, policy areas, and tools such as readiness and ethical impact assessments.
What it leaves out A summary, not the full Recommendation text; non-binding ethics guidance. Speaks to: Safeguards, Power, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
about 2.5 hrs Beginner Free
What you’ll get A practical, sector-neutral guide for organisations deploying AI: internal governance, human involvement, operations management and stakeholder communication.
What it leaves out Voluntary; written before generative and agentic AI. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
What you’ll get The shared baseline vocabulary most national AI policies borrow: five values-based principles plus five recommendations for governments.
What it leaves out High-level and non-binding; no enforcement or detailed obligations. Speaks to: Institutions, Safeguards, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
about 30 min (method page plus one inquiry) Beginner Free
What you’ll get Data labellers and content moderators researching and reporting on their own working conditions, in reports, zines and films.
What it leaves out First-hand accounts and case studies rather than policy proposals. Speaks to: Power, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Labour voice · International
about 1 hr (key findings) Intermediate Free
What you’ll get Evidence from 135 countries on the gap between responsible-AI commitments and implementation, with Global South versus North comparisons.
What it leaves out Measures government frameworks and evidence of action, not company practice. Speaks to: Institutions, Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global South
about 1 hr (executive summary and one country profile) Intermediate Free
What you’ll get Country-by-country scores for 19 Latin American and Caribbean countries, including a governance dimension covering institutions, international ties and regulation.
What it leaves out In Spanish. Measures national capacity and policy, not impacts on affected people. Speaks to: Institutions, Infrastructures, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
What you’ll get Plain-language start: AI as a question of power for the Global South, the global policy map, and questions to ask of AI systems.
What it leaves out A short primer centred on Africa; it maps debates rather than analysing them in depth. Speaks to: Power, Institutions, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global South
What you’ll get The global union movement's position on AI: bargaining rights, limits on surveillance, shared gains, and the Global South data-work supply chain.
What it leaves out A policy briefing for unions; not a legal analysis of any one country. Speaks to: Power, Safeguards, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Labour voice · International
45 min (summary) Intermediate Free
What you’ll get A sharp argument that AI policy is really about the concentration of power in a few tech firms, with a roadmap for action.
What it leaves out US-focused and openly argumentative; little on Global South policy. Speaks to: Power, Incentives, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
about 3.5 hrs Intermediate Free
What you’ll get A concrete example of an African national strategy: infrastructure, data and research pillars, plus governance, talent, ethics and local-language NLP.
What it leaves out A government plan of intentions; it does not itself create binding rules. Speaks to: Infrastructures, Institutions, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
What you’ll get Survey evidence from 206 content moderators on the human labour behind AI, with eight protocols they demand from tech firms.
What it leaves out Focuses on content moderation work, not AI governance more broadly. Speaks to: Safeguards, Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Labour voice · International
about 3 hrs (1 hr for executive summary and governance section) Intermediate Free
What you’ll get The continent-wide reference point: Africa's own priorities for AI, from governance and risk to infrastructure, skills, labour and investment.
What it leaves out A strategy, not binding law; national implementation is left to member states. Speaks to: Institutions, Infrastructures, Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
What you’ll get Latin America and the Caribbean's shared regional agenda, with a roadmap covering governance, future of work, vulnerable groups, environment and infrastructure.
What it leaves out A political declaration; the separate Roadmap holds the detail. Does not mention civil society. Speaks to: Institutions, Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
about 1 hr (final considerations chapter) Intermediate Free
What you’ll get A map of AI strategies, soft law and bills in eight Latin American countries, ending with human-rights-centred regulatory considerations.
What it leaves out In Spanish. Snapshot as of early 2024; bills have since moved. Speaks to: Institutions, Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · International
about 2 hrs (30 min for the job impact assessment tool) Intermediate Free
What you’ll get A job-impact checklist of opportunity and risk signals, plus practices for companies that build or deploy AI, shaped with frontline workers.
What it leaves out Voluntary guidance; aimed mainly at companies rather than regulators. Speaks to: Incentives, Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
about 2 hrs Intermediate Free
What you’ll get An evidence-based look at who actually got a say when Brazil, Chile, Colombia and Uruguay wrote their AI strategies.
What it leaves out Four countries, assessed as of 2022; focuses on consultation process, not strategy content. Speaks to: Power, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global South
What you’ll get A short, readable national policy that pairs adoption goals with a Responsible AI Office and practical ethics guidelines.
What it leaves out Brief policy; detailed rules sit in separate RURA ethics guidelines. Speaks to: Institutions, Infrastructures, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
about 2.5 hrs Intermediate Free
What you’ll get A practical template for African governments drafting a national AI strategy, built on five pillars and African use cases.
What it leaves out Written in 2021 for strategy designers; predates generative AI and the AU strategy. Speaks to: Infrastructures, Institutions, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
about 1.5 hrs Intermediate Free
What you’ll get A ten-part checklist for judging whether an algorithmic impact assessment gives affected people real consultation, access and redress.
What it leaves out US-centred and conceptual; not a ready-made assessment template. Speaks to: Safeguards, Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
about 10 min (short resolution) Beginner Free
What you’ll get Africa's main human-rights body asking states to keep AI under meaningful human control and grounded in African values and the African Charter.
What it leaves out A short call to action; detailed guidelines were left to a later study. Speaks to: Safeguards, Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global South
about 1.5 hrs Beginner Free
What you’ll get A well-documented case of an algorithm harming tens of thousands of families, and the safeguards that would have prevented it.
What it leaves out One country's public-sector case; written before the EU AI Act was adopted. Speaks to: Safeguards, Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
What you’ll get What workers can actually negotiate over hiring, workplace and performance algorithms: human appeal, joint audits, data limits and shared gains.
What it leaves out Workplace-level bargaining guide; predates generative AI. Speaks to: Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Labour voice · International
What you’ll get How an African grassroots community built language technology for its own languages, and why participation changes who AI serves.
What it leaves out A technical NLP paper; governance lessons are implicit. Speaks to: Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Affected-community voice · Global South
about 1 hr (introduction plus two country reports) Beginner Free
What you’ll get Civil-society authors describing how AI is actually used and governed in their own countries, mostly across the Global South.
What it leaves out Written in 2019, before generative AI; country reports vary in depth. Speaks to: Power, Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global South
What you’ll get Uses complex-systems ideas (emergence, feedback loops, cascading failures) to argue for when and how regulators should act under deep uncertainty.
What it leaves out Conceptual and principle-level; no worked regulatory design for a specific country. Speaks to: Timing, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
30 min (introduction and taxonomy); several hours in full Advanced Free
What you’ll get A taxonomy of where technical work can support governance: assessment, access, verification, security, operationalization and ecosystem monitoring.
What it leaves out Written for technical researchers and funders; catalogue described as incomplete. Speaks to: Safeguards, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
20 min (abstract and introduction); 4+ hrs in full Intermediate Free
What you’ll get Why compute (chips, data centres) is a practical lever for governing AI, and the privacy and power risks of using it.
What it leaves out Concentrates on frontier-model compute; less on data, labour or deployment harms. Speaks to: Infrastructures, Power, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 2 hrs Intermediate Free
What you’ll get A review of 68 articles that sorts AI governance research into technology, stakeholders and context, regulation, and processes.
What it leaves out Reviews academic literature, mostly organisational; not a practical how-to. Speaks to: Institutions, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
15 min (executive summary); about 1.5 hrs for the main text Intermediate Free
What you’ll get A glossary-style survey of how people define advanced AI and the governance tools, paradigms and theories of change around it.
What it leaves out Focused on advanced or frontier AI, not everyday algorithmic systems. Speaks to: Institutions, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get An early, readable account of the costs of ever-larger language models: environmental, financial, and harms from unexamined training data.
What it leaves out Written before ChatGPT-era products; focuses on language models specifically. Speaks to: Power, Safeguards, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 1.5 hrs Beginner Free
What you’ll get An African scholar's argument that imported AI and data business models can repeat colonial patterns, and a call for locally rooted AI.
What it leaves out An argumentative essay; not an empirical survey of African AI policy. Speaks to: Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global South
What you’ll get An overview of how AI is being deployed across Africa, its benefits and challenges, and the policy questions it raises.
What it leaves out Written in 2020, before the AU strategy and most national AI strategies. Speaks to: Infrastructures, Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global South
about 1 hr Intermediate Free
What you’ll get A step-by-step internal audit process (scoping, mapping, artifact collection, testing, reflection) that teams can adapt.
What it leaves out Internal company audits only; external and regulatory audits are outside its scope. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Industry voice · Global North
about 2 hrs Intermediate Free
What you’ll get Explains how colonial power patterns show up in AI and proposes concrete tactics for a decolonial AI practice.
What it leaves out Theory-led; offers tactics rather than specific regulatory measures. Speaks to: Power, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
Varies Beginner Partly free
What you’ll get A clear real-world case: a US health algorithm used spending as a proxy for need and so under-flagged Black patients.
What it leaves out One US healthcare algorithm; a case study, not a governance framework. Speaks to: Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 3 hrs; 20 min (introduction and overview) Intermediate Free
What you’ll get An early map of the field, splitting AI governance research into the technical landscape, AI politics, and ideal governance.
What it leaves out A 2018 research agenda; it predates current laws and generative AI. Speaks to: Institutions, Power, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get Evidence that commercial face-analysis systems failed most on darker-skinned women, and why testing by subgroup matters.
What it leaves out Tests gender classification by three vendors in 2017-18; systems have since changed. Speaks to: Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get A simple three-layer picture of AI governance: technical (algorithms and data), ethical principles, and social and legal institutions.
What it leaves out A short conceptual piece from 2017; no case studies or country detail. Speaks to: Institutions, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 1 hr Intermediate Free
What you’ll get Four habits for steering new technology while it can still change: anticipation, reflexivity, inclusion and responsiveness.
What it leaves out About emerging technology in general, not AI specifically. Speaks to: Timing, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 496 pages; 12-15 hrs Beginner Paid
What you’ll get Inside OpenAI's rise and the chips, data, cheap labour, energy and water frontier AI consumes, from Kenyan data workers to Chilean water activists.
What it leaves out Centred on one company and its leadership; not a guide to governance instruments. Speaks to: Power, Infrastructures, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Journalist voice · Global North
about 360 pages; 9-11 hrs Beginner Paid
What you’ll get A practical guide to telling working AI from hype, especially predictive AI used in hiring, schools, medicine, banking and justice.
What it leaves out Argues AI is not an existential risk; readers wanting the safety case should pair it with Russell. Speaks to: Safeguards, Power, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 320 pages; 8-9 hrs Beginner Paid
What you’ll get Stories of ordinary people affected by AI worldwide, including data workers, refugees and gig workers, showing the human side of automated systems.
What it leaves out Case studies rather than policy analysis; few concrete governance proposals. Speaks to: Power, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Journalist voice · Global North
about 224 pages; 6-7 hrs Beginner Paid
What you’ll get A readable case that Big Tech's data extraction repeats colonial patterns, with examples of workers and communities resisting it.
What it leaves out A critical argument about data and power; little on specific AI laws or technical safeguards. Speaks to: Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get A long history arguing that who benefits from new technology depends on power and choices, not on technology itself.
What it leaves out Economic history across many technologies; AI is one part of the argument. Speaks to: Power, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 352 pages; 9-10 hrs Beginner Paid
What you’ll get An industry insider's argument that AI and synthetic biology will spread fast and must be contained, and what containment could involve.
What it leaves out Written by a sitting AI company leader; a view from inside the industry. Speaks to: Timing, Power, Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Industry voice · Global North
about 224 pages; 6-7 hrs Beginner Paid
What you’ll get Why AI is being built on an unequal global order, what that means for poorer countries, and how production could be decentralised.
What it leaves out An argument about global inequality, not a guide to specific laws or technical controls. Speaks to: Power, Infrastructures, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · International
about 608 pages; 15-18 hrs Intermediate Paid
What you’ll get The clearest map of the three rival rulebooks for tech, US market-driven, Chinese state-driven, EU rights-driven, and how they compete worldwide.
What it leaves out Covers data, platforms and the internet broadly; AI is one part of a larger picture. Speaks to: Institutions, Power, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 248 pages; 6-8 hrs Beginner Paid
What you’ll get Why bias in facial recognition, lending and medical tools is built in by design, not a bug, and why algorithms must be held accountable.
What it leaves out US examples and US civil-rights framing. Speaks to: Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 216 pages; 6-7 hrs (chapters stand alone) Intermediate Free
What you’ll get African perspectives on responsible AI: epistemic justice, Ubuntu, AI at work in Kenya, gendered chatbots in Nigeria, and AI policy in North Africa.
What it leaves out An edited academic collection; chapters vary in depth and country coverage. Speaks to: Power, Institutions, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · International
about 336 pages; 8-10 hrs Beginner Paid
What you’ll get A first-person account of exposing bias in big-tech face recognition and founding the Algorithmic Justice League, written for non-specialists.
What it leaves out A memoir centred on face recognition and US advocacy rather than a survey of governance tools. Speaks to: Safeguards, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get Treats AI as a physical industry built on mining, energy, hidden labour and data, and asks who gains power from it.
What it leaves out A critical essay, not a policy toolkit. Speaks to: Power, Infrastructures.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 344 pages; 9-11 hrs Intermediate Paid
What you’ll get Four plain rules for AI law: support rather than replace experts, never pass as human, no arms races, always disclose who controls it.
What it leaves out A legal and normative argument, mostly US examples; not a guide to any current statute. Speaks to: Safeguards, Institutions, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 8-12 hrs (page count not confirmed) Beginner Paid
What you’ll get An accessible tour of what goes wrong when machine-learning systems do not do what we intend, and the researchers trying to fix it.
What it leaves out About the technical research problem; says little about laws, regulators or institutions. Speaks to: Safeguards, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Journalist voice · Global North
about 288 pages; 7-9 hrs Beginner Paid
What you’ll get The hidden human workforce, paid per task, that labels data and moderates content so AI appears automatic, and what protections they need.
What it leaves out Fieldwork predates generative AI; focused on platform labour rather than AI regulation. Speaks to: Power, Infrastructures, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 352 pages; 9-10 hrs Intermediate Paid
What you’ll get A leading AI scientist's case that machines more capable than us need rebuilding so they stay uncertain about, and defer to, human preferences.
What it leaves out Centred on long-run control of advanced AI; less on today's bias, labour or regulation. Speaks to: Safeguards, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get Shows how automated systems that look neutral can hide and deepen discrimination, and offers tools for resisting it.
What it leaves out US-centred; a critical framework rather than technical or policy steps. Speaks to: Power, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get How search engines can reproduce racist and sexist stereotypes, especially about Black women, and why that is a governance issue.
What it leaves out Centres on US search engines; not specific to newer AI systems. Speaks to: Power, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get Reported case studies of automated welfare, homelessness and child-protection systems in the US and what they did to poor families.
What it leaves out US public services only; written before generative AI. Speaks to: Power, Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 216 pages; 5-6 hrs Beginner Paid
What you’ll get A Kenyan view of how digital platforms reshape politics, from online activism and feminist mobilisation to misinformation in the 2017 elections.
What it leaves out About the internet and politics in Kenya, not AI specifically. Speaks to: Power, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Journalist voice · Global South
about 288 pages; 7-8 hrs Beginner Paid
What you’ll get The classic introduction to how opaque, unaccountable scoring models in hiring, credit, schools and insurance harm people and resist challenge.
What it leaves out Written in 2016, before generative AI; US cases. Speaks to: Safeguards, Power, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
15 min to explore Beginner Free
What you’ll get Searchable record of real-world AI failures and harms, useful for grounding abstract risk talk in concrete cases.
What it leaves out Built from collected public reports; the about page shows no current incident total. Speaks to: Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
15 min to explore Beginner Free
What you’ll get A near-real-time, news-based feed of AI incidents and hazards worldwide, using OECD's common definitions.
What it leaves out Built from news reports, so it captures only part of all incidents; entries are machine-classified. Speaks to: Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
30 min to understand; longer to complete Beginner Free
What you’ll get A working, reusable questionnaire that scores an automated decision system into one of four impact levels.
What it leaves out Designed for Canadian federal departments under one directive. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
20 min to explore Intermediate Free
What you’ll get Open, frequently updated data on AI models, training compute, chips, data centres and companies - the numbers behind compute-governance debates.
What it leaves out Quantitative trends only; no policy analysis or legal coverage. Speaks to: Infrastructures, Incentives, Power.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
6 days intensive or 6 weeks part-time Intermediate Free
What you’ll get A structured, discussion-based tour of frontier AI evidence, institutions, power and competing governance strategies, ending with a personal action plan.
What it leaves out Focused on frontier/advanced AI; cohort-based with application, not self-paced. Speaks to: Institutions, Power, Timing, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
about 3 hrs Intermediate Partly free
What you’ll get A short university course on putting governance and responsible-AI practice around generative AI inside an organisation.
What it leaves out Organisational focus; certificate requires payment or Coursera Plus. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
What you’ll get Plain-language explainer of what an AI safety/security institute does, what it cannot do, and where it fits in governance.
What it leaves out UK-specific; one institution only. Speaks to: Institutions, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Civil-society voice · Global North
15 min to explore Beginner Free
What you’ll get Look up what any of ~70 governments has actually adopted on AI policy and strategy, in one comparable place.
What it leaves out Lists initiatives as governments report them; the page gives no last-updated date. Speaks to: Institutions, Incentives.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · International
1 hr (Policy and Governance chapter and top takeaways) Beginner Free
What you’ll get The yearly data reference for AI research, performance, economy, public opinion and, in Chapter 8, policy and governance trends.
What it leaves out A data compendium, not policy advice; very long in full. Speaks to: Incentives, Infrastructures, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
20 min to explore Beginner Free
What you’ll get A ready-made template and guidance for public bodies to publish what algorithmic tools they use and how.
What it leaves out Mandatory only for UK central government and some arm's-length bodies. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
15 min (factsheet) Intermediate Free
What you’ll get A government evaluator's data-driven summary of how frontier AI capabilities have evolved, drawn from two years of testing.
What it leaves out Focused on frontier capabilities relevant to national security and public safety. Speaks to: Safeguards, Institutions, Timing.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Government voice · Global North
20 min to explore Beginner Free
What you’ll get One searchable map of AI risks gathered from dozens of frameworks, sorted by cause and domain.
What it leaves out Catalogues risks; does not rank them or prescribe responses. Speaks to: Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
20 min to explore Intermediate Free
What you’ll get Read and compare the text of AI laws and governance documents, with summaries and tags for risks and strategies.
What it leaves out Covers 'hundreds' of documents, not every jurisdiction; some summaries await peer review. Speaks to: Institutions, Safeguards.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North
about 1-2 hrs per workbook Beginner Free
What you’ll get Practical, workshop-ready workbooks for applying a process-based governance framework to public-sector AI projects, with case studies.
What it leaves out Built around UK public-sector practice; workbooks rather than a guided online course. Speaks to: Safeguards, Institutions.
Incentives Institutions Infrastructures Safeguards Timing Power Last checked 31 Oct 2026 · Research voice · Global North