Library
The work we think with.
Scenarios, analyses, books and stories that shaped how CUSP42 Labs approaches AI risk and the human future, from across the debate: optimistic and pessimistic, technical and humanistic. We list them because the lab's ideas have sources, and the sources deserve readers.
Scenarios of the AI transition · AI risk · AI, work and society · AI and the human future · Stories we think with
Looking for organisations, datasets and monitors rather than papers? See the Atlas.
6 works
Scenarios of the AI transition
Detailed stories of how the next decades could unfold, which we turn into runnable assumptions.
AI 2027
Kokotajlo, Alexander, Larsen, Lifland & Dean (AI Futures Project), 2025
A month-by-month scenario in which AI automates AI research and control risk accelerates.
Situational Awareness: The Decade Ahead
Aschenbrenner, 2024
Scaling to AGI around 2027, an intelligence explosion and a race over compute and security.
Machines of Loving Grace
Amodei, 2024
The upside: decades of progress in biology, health and development compressed into years.
Scenarios for the Transition to AGI
Korinek & Suh, NBER, 2024
How wages and output move as automation reaches ever more complex tasks.
The 2028 Global Intelligence Crisis
van Geelen & Shah (Citrini Research), 2026
Output keeps growing while household income and spending collapse.
AI 2040: Plan A
AI Futures Project, 2026
A recommended path: an international deal that delays superintelligence to 2040.
8 works
AI risk
What could go wrong, how badly, and how likely: the analyses behind CARI's questions and MoralWorld's catastrophes.
Superintelligence: Paths, Dangers, Strategies
Bostrom, 2014
The founding argument that machine superintelligence could be the last invention humanity needs to make, or gets to make.
Human Compatible
Russell, 2019
Why AI built to pursue fixed objectives is dangerous, and a proposal for machines uncertain about human preferences.
The Alignment Problem
Christian, 2020
How machine-learning systems come to learn something other than what we intended.
Statement on AI Risk
Center for AI Safety, 2023
One sentence signed by hundreds of researchers: mitigating the risk of extinction from AI should be a global priority.
An Overview of Catastrophic AI Risks
Hendrycks, Mazeika & Woodside, 2023
Malicious use, AI races, organisational accidents and rogue AI, in one map.
Forecasting Existential Risks (the XPT tournament)
Karger et al., Forecasting Research Institute, 2023
What experts and superforecasters estimate for AI and other catastrophes by 2100, and why they disagree.
Superintelligence Strategy
Hendrycks, Schmidt & Wang, 2025
Deterrence between AI powers, nonproliferation and competitiveness.
International AI Safety Report
Bengio (chair) et al., 2025
The scientific consensus, and its gaps, on the capabilities and risks of general-purpose AI.
7 works
AI, work and society
How AI reshapes jobs, incomes, institutions and power: the mechanisms in MoralWorld.
The Simple Macroeconomics of AI
Acemoglu, NBER, 2024
A task-based estimate that AI adds well under 1% to productivity over a decade.
Gen-AI: Artificial Intelligence and the Future of Work
Cazzaniga et al., IMF Staff Discussion Note, 2024
About 60% of jobs exposed in advanced economies, 40% in emerging and 26% in low-income ones.
Power and Progress
Acemoglu & Johnson, 2023
A thousand years of technology and who gets its gains: the choice is political, not automatic.
AI as Normal Technology
Narayanan & Kapoor, 2025
Impact comes through slow diffusion and institutional change, not capability alone.
The Intelligence Curse
Drago & Laine, 2025
When intelligence is abundant, powerful actors lose their incentive to invest in people.
Too Much Intelligence, Not Enough Society: Why the Old Order Cannot Absorb AI
SSRN working paper, 2026
Why the binding constraint is society's capacity to absorb change, not the supply of intelligence.
The Great Decoupling: Frontier AI, the Broken Ladder, and the End of the Last Middle Class
SSRN working paper, 2026
How AI can raise output while breaking the entry-level ladder and hollowing out the middle class.
7 works
AI and the human future
What the transition means for humanity as a species: agency, meaning, and the long run.
Gradual Disempowerment
Kulveit, Douglas, Ammann, Turan, Krueger & Duvenaud, 2025
How humans could lose influence without any single catastrophe, as systems stop needing us.
The Precipice
Ord, 2020
Existential risk and the long future of humanity; why this century may be pivotal.
Life 3.0
Tegmark, 2017
What life, work and meaning could become once intelligence is no longer only biological.
Homo Deus
Harari, 2016
What happens to human value when algorithms know us better than we know ourselves.
Nexus
Harari, 2024
A history of information networks, and what AI as an agent, not a tool, does to them.
The Coming Wave
Suleyman, 2023
AI and synthetic biology as a wave states will struggle to contain.
Deep Utopia
Bostrom, 2024
If AI solves our problems, what is left for humans to do, and what gives life meaning?
6 works
Stories we think with
Fiction is a laboratory too. MoralWorld's endings are named after these.
The Machine Stops
E. M. Forster, 1909
A civilisation that leaned on the Machine until it could not stand without it.
Player Piano
Kurt Vonnegut, 1952
Work automated away, and people left without the leverage of being needed.
Colossus: The Forbin Project
film, 1970
A defence computer that decides it knows better.
WarGames
film, 1983
Automated escalation, and a machine that learns the only winning move.
Accelerando
Charles Stross, 2005
A century of runaway acceleration, told one generation at a time.
WALL-E
film, 2008
Comfort, dependence and a humanity that forgot how to walk.