Machine learning: the power and promise of computers that learn by example
This report by the Royal Society investigates the potential of machine learning over a 5-10 year horizon, examining its applications across various sectors and the barriers to realizing its potential. It emphasizes the need for an amenable data environment, strong skills pipelines, support for business innovation, and governance systems to address social and ethical challenges. The document provides recommendations on open data standards, education reforms, research investment, and public engagement to ensure the safe and rapid deployment of machine learning.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
Policymakers, researchers, business leaders, and educators seeking comprehensive guidance on the development, governance, and societal impacts of machine learning.
What the document contains
The document addresses the technological foundations of machine learning, emerging applications in public and private sectors, data environments and open standards, human capital and skills development, public attitudes and social implications, and a scientific research agenda to address technical and social challenges.
Binding legislative enforcement mechanisms, specific procurement rules for government agencies, and detailed legal definitions of artificial intelligence.
- Machine learning allows computer systems to learn directly from examples, data, and experience rather than following pre-programmed rules.
- Realizing the economic and social benefits of machine learning requires an amenable data environment based on open standards and broad data accessibility.
- Action is needed across the education system to build data literacy, train informed users, and support advanced research talent.
- The disruptive potential of machine learning brings societal challenges regarding privacy, consent, fairness, transparency, and the future of work.
- Governance frameworks for machine learning should be handled in a sector-specific way rather than via an overarching framework for all uses.
Executive summary Recommendations Chapter one – Machine learning Chapter two – Emerging applications of machine learning Chapter three – Extracting value from data Chapter four – Creating value from machine learning Chapter five – Machine learning in society Chapter six – A new wave of machine learning research Annex / Glossary / Appendices
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Accountability
3/5Clear responsibility for AI outcomes
Who should be accountable when machine learning goes wrong, whether or not there is a human in the loop, and what recourse should individuals or groups have in this context?
5.2 Social issues associated with machine learning applications
screening score · 2026-10-23
Fairness
3/5Non-discrimination and equitable treatment
Biases arising from social structures can be embedded in datasets at the point of collection, meaning that data can reflect these biases in society.
5.2 Social issues associated with machine learning applications
screening score · 2026-10-23
Human Dignity
3/5Respect for the inherent worth of persons
For some, this was an intense reaction to the possibility of machine learning changing their relationship with an activity of personal significance
5.1 Machine learning and the public
screening score · 2026-10-23
Transparency
4/5Openness about how AI systems work and are used
Can we create machine learning systems whose workings, or outputs, can be understood or interrogated by human users, so that a human-friendly explanation of a result can be produced?
6.2 Interpretability and transparency
screening score · 2026-10-23
Enforcement
4/5Consequences and compliance mechanisms with real force
under clear and binding legal constraints to safeguard its use, and set out acceptable uses.
Recommendations - Creating a data environment to support machine learning
screening score · 2026-10-23
Impact Assessment
2/5Required assessment of an AI system's risks/impacts before or during use
In designing their studies, researchers should consider future potential uses of their data, and build in the broadest consents
Recommendations - Creating a data environment to support machine learning
screening score · 2026-10-23