GCRI · Artificial Intelligence and Autonomous Systems
Dependencies and consequences
AI systems depend on data, compute, energy, software supply chains and people, and they are increasingly embedded in services and infrastructures that others depend on. This page explains the causal, physical, economic, social and ecological links between AI and other systems, how failures travel along them and how the platform analyses them while stating what is uncertain. Concentration is the recurring theme.
- STEM methods
- 7
- Canonical methods this field applies
- Topics
- 8
- Subject pages in the field
- Counterparts
- 3
- Platforms in shared fields across the network
- Report series
- 2
- Reports series it contributes to
Upstream dependencies.
Chips, data centres and cloud
Training and running large models depends on concentrated chip supply and a few cloud providers. See Compute and energy.
Physical resources
Data centres draw electricity and, for cooling, often water, linking AI growth to grids and local water stress.
Inputs and labour
Training data, human feedback and annotation, often provided by low-paid workers, shape model behaviour.
Field overview.
The screen a team would use to follow this subject: its measures, states and decisions. Names are examples and values are illustrative.
Downstream consequences.
- 01
Public services
Automated decisions in benefits, health, policing or migration affect rights directly. See AI and public services.
- 02
Infrastructure and operations
AI in grid management, logistics and industrial control can propagate errors into physical systems. See Autonomous physical systems.
- 03
Security
AI changes the speed and scale of cyber attack and defence, and models are themselves targets. See AI and cyber security.
- 04
Finance
Shared models and data can make institutions behave alike under stress, a concern examined in AI and Financial Model Risk at The Global Risks Alliance (GRA).
- 05
Information and trust
Generated content at scale affects the information environment and trust in institutions.
Common modes and correlated errors.
Because many applications are built on the same foundation models, a flaw or vulnerability can appear simultaneously across sectors that otherwise share little. An update by one provider can change behaviour in thousands of downstream systems. Outages at a major cloud region can stop services that never knew they shared it.
Social effects travel more slowly but last longer. Errors concentrated on particular groups erode trust, and changes in work and skills reshape labour markets over years. See Labour and social effects.
How dependencies are analysed.
The platform maps dependencies using system analysis, supply-chain mapping and incident studies, with probability and statistics to express uncertainty and cybersecurity and cryptography for threat modelling. A dependency map is not a forecast: it shows how systems are connected and how failure could travel, and each link carries its evidence label. Governance responses are convened by AI and Digital Governance at The Global Risks Forum (GRF).
Questions and answers.
Why does concentration matter?
How significant is AI’s energy use?
Are specific providers’ weaknesses published?
Who governs these dependencies?
Contribute evidence from this page.
Reports is the Consortium’s quarterly publication. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work.
- EmailGateway GW-GCRISend it from this pageEvidence, a correction, a case, a method, a review, a translation or profile input. It arrives tagged with this page, so the editor who owns it receives it.
- ZenodoOpen depositDeposit it and submit to the communityPublish your own record on Zenodo with your ORCID iD, then submit it to the Reports community there. Curators review it; the DOI stays yours.
- ReportsGCRISee every way to contributeThe contribution types, how each is reviewed, what credit you receive, the licence and what is never accepted.
Never send personal data, sensitive infrastructure detail or anything you have no right to share. A contribution is reviewed before use; sending it confers no authorship or editorial rights, and sponsorship never buys a finding.
Bring a method, a dataset or a question.
Information and enquiries are open today. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.
Legal documents.
Maps, data and figures on this site are shown for context and general information. They are not advice, forecasts or warnings; the limits of the information are set out in the Disclaimers and Data Notices.