Designing Social Robots with Ethical, User-Adaptive Explainability in the Era of Foundation Models
This paper argues that ethical, user-adapted explainability must be treated as a core design objective for foundation-model-driven social robotics. The authors identify open challenges regarding pre-training biases, generic explanations, and opacity in adaptive systems, alongside ethical risks such as misalignment, epistemic exclusion, and institutional externalisation of responsibility. To address these issues, the paper proposes four concrete recommendations: designing modality- and user-aware explanation strategies, employing co-design, integrating layered adaptation, and prioritizing smaller, fairer datasets.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
Researchers, designers, and developers working at the intersection of human-robot interaction, artificial intelligence, and ethics should read this to understand the challenges and design principles for explainable social robots.
What the document contains
The paper addresses the intersection of foundation models, social robotics, user adaptation, and explainability. It maps out open technical and ethical challenges, and offers four recommendations for designing user-adapted, modality-aware, and co-designed explanation strategies.
The document does not provide definitive technical solutions or source code, nor does it establish binding regulatory enforcement mechanisms.
- Traditional one-size-fits-all explanation strategies are inadequate for foundation-model-driven social robots.
- Foundation models introduce ethical risks like pre-training bias, epistemic exclusion, and the commodification of ethical responsibility.
- Explanations should be modality- and user-aware, tailoring the content and channel to individual capabilities and preferences.
- Co-design processes involving end-users and professionals are essential for determining effective explanation needs and adaptation criteria.
- Layered adaptation and smaller, curated datasets should be prioritized over indiscriminate scaling to achieve fairer and more robust robotic interactions.
· 1 Introduction · 2 Related Work · 3 Open Challenges - 3.1 User-Adaptive Explainability - 3.2 Ethical Concerns · 4 Towards Realisation and Recommendations - 4.1 Illustrative Use Case · 5 Considerations and Concerns
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Human Oversight
0/5Humans remain in control of consequential decisions
screening score · 2026-10-23
Privacy
1/5Protection of personal data and private life
These strategies help avoid ethically fraught practices like over-personalisation, unjustified data harvesting, or reinforcing inequities
4 Towards Realisation and Recommendations
screening score · 2026-10-23
Enforcement
0/5Consequences and compliance mechanisms with real force
screening score · 2026-10-23