Fair contracts for artificial intelligence?
Clinical Snapshot
PICO Framework
| P — Population | Artificial intelligence systems and their alignment frameworks; by extension, developers, ethicists, and policymakers involved in AI governance |
| I — Intervention | Resource-rational contractualism (RRC) as a normative framework for AI alignment and moral cognition |
| C — Comparator | Standard philosophical contractualism and existing AI alignment approaches |
| O — Outcomes | Conceptual adequacy of RRC to capture contractualist moral principles; feasibility of guiding efficient AI moral cognition through impartial deliberation |
Bottom Line
This theoretical commentary from two Google DeepMind researchers proposes that a modified resource-rational contractualism (RRC), anchored in impartial deliberation, can serve as a normative framework for AI alignment and moral reasoning. While the question is intellectually important and the publication venue prestigious, this is a philosophical opinion piece with no empirical data, no patient outcomes, and no clinical validation. The most significant concern is a substantial conflict of interest: both authors are employed by a major commercial AI developer, which may shape which alignment frameworks are considered viable or desirable. For clinicians and health system leaders, the paper's value lies in its contribution to the theoretical foundations of AI ethics rather than any immediate practice change. Before any alignment framework derived from RRC could be applied to clinical AI systems in Australia, it would require operationalisation, independent validation, and scrutiny against TGA SaMD requirements and NHMRC ethical standards. Senior clinicians should treat this as a horizon-scanning piece relevant to AI governance discussions, not as actionable clinical guidance.
Key Findings
Effect Size: Not applicable — no quantitative effect size reported
Primary Outcome: A modified version of resource-rational contractualism (RRC), centred on impartial deliberation, is proposed as a viable framework for identifying AI alignment principles and guiding efficient AI moral cognition
Nnt Or Sensitivity: Not applicable — no clinical or diagnostic metrics reported
Confidence Interval: Not applicable — theoretical commentary with no statistical analysis
Clinical Application
The theoretical framework proposed requires substantial operationalisation before clinical application. No implementation pathway, validation methodology, or practical guidance for healthcare settings is provided. Adoption would require interdisciplinary collaboration between ethicists, clinicians, regulators, and AI engineers. In Australia, the Therapeutic Goods Administration (TGA) regulates software as a medical device (SaMD), including AI-enabled clinical tools, under the Medical Device Regulations 2002. The Australian Digital Health Agency (ADHA) and the RACGP have published guidance on responsible AI use in general practice. The ethical framework proposed in this commentary could theoretically inform TGA post-market surveillance requirements and RACGP AI governance standards, but no direct translation is provided. The National Health and Medical Research Council (NHMRC) ethical guidelines for AI in health would be the relevant Australian reference point. PBS and TGA approval processes for AI-driven therapeutics remain nascent, and a principled alignment framework could contribute to regulatory development. This commentary is not directly applicable to patient populations. It is relevant to AI developers, clinical informaticists, health technology assessment bodies, bioethicists, and policymakers involved in the governance of AI-enabled clinical decision support systems
Abstract
This comment explores resource-rational contractualism (RRC) as a framework for artificial intelligence (AI) alignment and reasoning. It assesses RRC's capacity to capture the insights of philosophical contractualism and canvasses its extension to AI alignment. It suggests that a modified version of RRC, centered on impartial deliberation, can be used to identify principles for alignment and to guide efficient AI moral cognition.
References
- 1.Gabriel, I., & Franklin, M. (2026). Fair contracts for artificial intelligence? Behavioral and Brain Sciences. https://doi.org/10.1017/S0140525X25101714
Related Research
Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association
Artificial Intelligence Tools for Gastrointestinal Research: A Practical Guide
2 Aug 2026
Clinical spine surgery
An Introduction to Machine Learning for the Practicing Spine Surgeon
2 Aug 2026
Rheumatic diseases clinics of North America
Sources of Bias in Clinical Artificial Intelligence and Applications in Rheumatology
2 Aug 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service