Project Cassandra

The final part broadens the inquiry beyond individual, institutional and national concerns to consider the wider consequences of increasingly capable artificial intelligence for society and humanity as a whole.

Building on the human, social and governance challenges explored in the preceding three parts, Systemic and Global Realities examines the interconnected risks, global implications and longer-term uncertainties that may arise as AI capabilities develop. It considers how developments in one area may have consequences elsewhere, potentially extending beyond the reach of existing institutions and national boundaries.

The emphasis remains on preparation, responsibility and the choices humanity may need to make in shaping its future. These wider considerations bring together the principal themes of Project Cassandra and lead towards its concluding question: whether humanity can establish effective governance before increasingly autonomous AI moves beyond its meaningful control.

When AI Capability Becomes Concentrated - Examines the implications of increasingly powerful AI capabilities becoming concentrated within a small number of organisations or countries. It considers how such concentration could influence economic opportunity, political power and decision-making, potentially creating new dependencies and inequalities that extend beyond national borders.

Measuring Whether People Actually Benefit - Examines how societies might determine whether advances in AI are genuinely improving people's lives, rather than simply increasing productivity or economic growth. It considers the need to assess benefits in terms of economic security, wellbeing, opportunity and quality of life, ensuring that technological progress translates into meaningful improvements for society.

Ensuring the Benefits Reach Beyond National Borders - Explores how the benefits of increasingly capable AI might be shared beyond the countries and organisations leading its development. It considers the risk of widening global inequalities and the need for international cooperation to ensure that less technologically advanced nations can participate in, and benefit from, the opportunities AI creates.

Reaching People Whose Governments Do Not Cooperate - Examines the difficulties of ensuring that people benefit from international AI cooperation when their governments are unwilling or unable to participate. It considers how international arrangements might reach and support affected populations without depending entirely on the cooperation of their national authorities.

The Environmental and Resource Cost of AI - Examines the energy, water, infrastructure and other resources required to develop and operate increasingly capable AI systems. It considers the environmental pressures arising from their expansion, alongside AI's potential to improve resource efficiency and help address environmental challenges. It also explores the need to assess these costs and benefits together.

Resilience When Society Depends Upon AI - Examines how society can maintain essential services, infrastructure and decision-making as dependence on AI systems increases. It considers the risks of failures, disruption or loss of access to critical AI capabilities, and the need for contingency arrangements that preserve society's ability to function when those systems cannot be relied upon.

Knowing What We Know - Examines the importance of distinguishing established evidence from assumptions, forecasts and speculation when assessing AI's possible future consequences. It considers how uncertainty should be acknowledged, claims critically examined and decisions grounded in the best available evidence, whilst recognising that preparation cannot always wait for certainty.

What the Research Has Established - Examines what existing research and evidence can reasonably establish about AI's development and its implications for humanity. It considers the findings that support concern or optimism, identifies significant gaps in current knowledge and distinguishes conclusions supported by evidence from questions that remain unresolved.

Knowing Where Cassandra Stops - Clarifies the scope and limitations of Project Cassandra's inquiry, distinguishing its evidence-based concerns and proposals from predictions or definitive conclusions. It recognises that the practical arrangements explored throughout the book require further investigation, specialist expertise and democratic consideration, rather than presenting them as a complete political programme.

The Direction of Travel - Draws together the principal themes of Project Cassandra, considering how humanity might approach an uncertain AI future with preparation, responsibility and informed choices. It emphasises the importance of acting before circumstances force decisions upon us, whilst recognising the opportunities that increasingly capable AI may offer.