Kulveit, J., Douglas, R., Ammann, N., Turan, D., Duvenaud, D. & Krueger, D. (2025). Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development. arXiv:2501.16946. Read

Preprint · argument

The core risk: that human participation has kept economic, cultural and state systems aligned with human interests, and that displacing it erodes that alignment without any single dramatic event.

Jha, T., Everitt, T. & Grzankowski, A. (2026). Human Amplification, Intelligent Agents, and the Aims of AI Research. IASEAI 2026. Read

Conference paper · argument

The proposition behind “amplify, not replace”: that the primary objective of AI research should be systems that augment human abilities and preserve human agency.

Jha, T., Svarc, R. & Everitt, T. (2026). Towards Characterising Conditions of Assistance, Hindrance and Depletion in Human Agency from AI Systems. Manuscript.

Manuscript · argument

Four dimensions of agency, the distinction between assistance, hindrance and depletion, and five maxims for artificial assistants. Directly relevant to our assessment framework.

Seligman, M. E. P. (2026). Agency: The Psychological History of Human Progress.

Book · argument

Agency as efficacy, optimism and imagination; the account we set our four pillars against.

Kosmyna, N., Hauptmann, E., Yuan, Y. T., Situ, J., Liao, X.-H., Beresnitzky, A. V., Braunstein, I. & Maes, P. (2025). Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. MIT Media Lab. arXiv:2506.08872. Read

Preprint · early evidence · contested

EEG connectivity was strongest in the tool-free group and weakest in the ChatGPT group, with lower recall and lower ownership of the work. Small sample (54; 18 in the final session); the authors call it preliminary, and a published critique (Stanković et al., 2026, arXiv:2601.00856) questions the methodology. We cite it as early evidence, not proof.

Maguire, E. A. et al. (2000). Navigation-related structural change in the hippocampi of taxi drivers. PNAS.

Peer-reviewed · evidence

The London Knowledge: the practice of navigation measurably reshapes the brain.

Dahmani, L. & Bohbot, V. D. (2020). Habitual use of GPS negatively impacts spatial memory during self-guided navigation. Scientific Reports.

Peer-reviewed · evidence

Satnav use is associated with poorer spatial memory over time.

Sparrow, B., Liu, J. & Wegner, D. M. (2011). Google Effects on Memory. Science.

Peer-reviewed · evidence

We come to remember where to find information rather than the information itself.

Risko, E. F. & Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences.

Peer-reviewed review · evidence

The review that names and frames the phenomenon.

Lee, H.-P. et al. (2025). The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects from a Survey of Knowledge Workers. CHI 2025.

Peer-reviewed · self-report evidence

Greater confidence in AI is associated with less critical engagement. Survey-based.

Parasuraman, R. & Manzey, D. H. (2010). Complacency and Bias in Human Use of Automation: An Attentional Integration. Human Factors.

Peer-reviewed review · evidence

People tend to trust the machine even when a human is checking it.

Skitka, L. J., Mosier, K. L. & Burdick, M. (1999). Does automation bias decision-making? International Journal of Human-Computer Studies.

Peer-reviewed · evidence

The original automation-bias experiments.

Fogg, B. J. (2003). Persuasive Technology: Using Computers to Change What We Think and Do.

Book · argument

The foundational text on persuasive design.

Williams, J. (2018). Stand Out of Our Light: Freedom and Resistance in the Attention Economy. Cambridge University Press.

Book · argument

Attention as the ground of agency.

Zuboff, S. (2019). The Age of Surveillance Capitalism.

Book · argument

The commercial incentives that reward dependence.

Bandura, A. (2001). Social Cognitive Theory: An Agentic Perspective. Annual Review of Psychology.

Peer-reviewed review · argument

The mainstream psychological account of agency and efficacy.

Sen, A. (1999). Development as Freedom.

Book · argument

Agency and capability; and why, for much of the world, AI may add agency rather than erode it.

Mackenzie, C. & Stoljar, N. (eds.) (2000). Relational Autonomy.

Book · argument

The philosophical basis of our relational pillar: no one’s agency is exercised alone.

Clark, A. & Chalmers, D. (1998). The Extended Mind. Analysis.

Peer-reviewed · argument

The idea invoked in our discussion of brain-computer interfaces.

Illich, I. (1973). Tools for Conviviality; and (1974) Energy and Equity.

Books · argument

The ancestor of our bicycle and car: tools that extend human capacity, and tools that past a threshold consume it.

McLuhan, M. (1964). Understanding Media.

Book · argument

Every technology as both extension and amputation.

Wittgenstein, L. (1953). Philosophical Investigations, §107.

Book · argument

“Back to the rough ground”: the source of the smooth-ice image in our lens.

Indongo, N. N., Smith, R. C., Winschiers-Theophilus, H., Kandjengo, S. & Mbinge, U. (2026). Designing Responsible AI with Indigenous Communities: Towards Participatory AI with ovaHimba and San People in Namibia. Proceedings of the 19th Participatory Design Conference (PDC ’26), 446–456. Read

Peer-reviewed · evidence and practice

Elders leading, communities controlling their own data, people paid for their time, local people trained to run the systems: the model for how we work.

Lewis, J. E., Arista, N., Pechawis, A. & Kite, S. (2018). Making Kin with the Machines. Journal of Design and Science, MIT.

Peer-reviewed · argument

Hawaiian, Cree and Lakota perspectives on relationship with AI.

Lewis, J. E. (ed.) (2020). Indigenous Protocol and Artificial Intelligence Position Paper.

Position paper · argument

Includes Suzanne Kite’s account of the full life cycle of an AI system, stone and metal included, through Lakota ontology.

Carroll, S. R. et al. (2020). The CARE Principles for Indigenous Data Governance. Data Science Journal.

Peer-reviewed · framework

Collective benefit, authority to control, responsibility, ethics.

UNESCO (2023). Guidance for generative AI in education and research.

Policy

The international reference for AI in schools, including age limits and human oversight.

Odgers, C. L. (2024). The great rewiring: is social media really behind an epidemic of teenage mental illness? Nature, 628, 29–30.

Book review · caution

The evidence linking screens to children’s mental-health decline is weaker than public debate assumes. Our education work will follow the evidence.

European Union (2024). Artificial Intelligence Act, Article 27: Fundamental Rights Impact Assessment.

Law · framework

The model for an agency impact assessment.

Council of Europe (2024). HUDERIA: Methodology for the Risk and Impact Assessment of AI Systems from the Point of View of Human Rights, Democracy and the Rule of Law.

Policy · framework

A procedural model for validation and governance.

Gordon, E. C. & Seth, A. K. (2024). Ethical considerations for the use of brain–computer interfaces for cognitive enhancement. PLOS Biology, 22(10), e3002899.

Peer-reviewed · argument

Includes the concern we call mental monoculture: the narrowing of thought toward a single pattern.

Laukkonen, R., Krier, S., Bakalar, C., et al. (2026). Positive Alignment: Artificial Intelligence for Human Flourishing. arXiv:2605.10310. Read

Preprint · argument

AI should actively support human flourishing, not only avoid harm.