GCRI · Artificial Intelligence and Autonomous Systems
Field overview
AI and autonomous systems as a risk field studies how learned models, software agents and robots behave in the real world, how their behaviour can be measured and how their failures propagate into the institutions, infrastructures and lives that depend on them. This page defines the field, explains why it matters for society, sets out what experts study and shows how it connects to other risks.
- 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
From algorithms to deployed systems.
Models and data
How training data, objectives and architectures shape behaviour, including failure modes.
Evaluation and oversight
How systems are tested, monitored and kept under meaningful human control.
Autonomy in the world
How sensing, planning and control interact in robots and vehicles.
The field combines artificial intelligence and machine learning, the methods by which systems learn from data, with AI evaluation and assurance engineering, the engineering discipline of testing and assuring them. It extends to robotics and automation, where AI controls physical machines, and to high-performance, cloud and edge computing, the infrastructure that trains and runs models.
As a risk field it is defined by consequences: what happens when a model generalises badly, when an agent misuses a tool, when a robot misreads its environment, or when many systems share the same flaw. social and behavioural science is part of the field because the effects of AI depend on how people use, trust and contest it.
Field overview.
The screen a team would use to follow this subject: its measures, states and decisions. Names are examples and values are illustrative.
AI is becoming infrastructure.
- AI is moving from advisory tools to systems that act, through software agents and physical autonomy.
- Public services increasingly use automated systems in decisions about people’s rights and access.
- A small number of model providers, cloud platforms and chip supply chains underpin many applications.
- Evaluation science lags capability, so claims about safety and performance are often hard to check.
The core questions.
- 01
Measurement
Do evaluations measure what they claim, and do results generalise? See Model evaluation.
- 02
Action
How should authority be delegated to systems that act? See Agentic systems and Autonomous physical systems.
- 03
Security
How can models be attacked, and how can AI change cyber offence and defence? See AI and cyber security.
- 04
Use
How does AI change science and public services? See AI in science and AI and public services.
- 05
Foundations and effects
What does AI depend on, and whom does it affect? See Compute and energy and Labour and social effects.
How AI touches other risks.
AI connects to cybersecurity, energy, labour markets, finance and public trust. Dependencies and consequences explains those links. Governance is convened through AI and Digital Governance at The Global Risks Forum (GRF), and financial-sector model risk is studied through AI and Financial Model Risk at The Global Risks Alliance (GRA).
Pages in Field overview.
Questions and answers.
Is this field the same as AI ethics?
Which methods does the field rely on?
Does the field cover military AI?
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.