CEIAI Governance Database
Back to the database
Record STMT-W2531F6Aspirational

Report from Dagstuhl Seminar 20021 Spoken Language Interaction with Virtual Agents and Robots (SLIVAR): Towards Effective and Ethical Interaction

Report from Dagstuhl Seminar 20021 Spoken Language Interaction with Virtual Agents and Robots (SLIVAR): Towards Effective and Ethical Interaction

Report International lawGlobaladopted

This report documents the outcomes of Dagstuhl Seminar 20021 on Spoken Language Interaction with Virtual Agents and Robots (SLIVAR), held in January 2020. It brings together world experts on spoken language processing and human-robot interaction to review open research questions, propose evaluation protocols, investigate data sharing, and examine ethical and societal issues. The seminar highlights the necessity of aligning visual, vocal, behavioral, and cognitive affordances to avoid the uncanny valley and bridge the habitability gap in human-machine interactions.

Summary and content fields are drawn from the document text and are pending editorial review.

Who issued it

Schloss Dagstuhl – Leibniz-Zentrum f%fc r InformatikAcademic
Publisher
Legal forceAspirational
Under whose authorityInternational law
Judged fromfull text

Who it governs

Researchers and academicsAI developers and providersAI deployers and users

This report is for researchers, academics, and developers working at the intersection of spoken language processing, robotics, and artificial intelligence ethics.

Sectors
Research and science · artificial intelligence and robotics research

What the document contains

What's covered

The document addresses the integration of spoken language processing and human-robot interaction, technical architectures for multimodal systems, evaluation protocols, ethical implications of affective computing and social robots, and open research problems in dialogue management.

What's missing

It contains no legally binding regulations, statutory compliance mandates, or enforceable enforcement mechanisms for AI deployment.

Key takeaways
  • Spoken language processing and human-robot interaction communities need closer collaboration to establish a distinct SLIVAR research community.
  • The visual, vocal, behavioral, and cognitive affordances of artificial agents must be aligned to avoid the uncanny valley effect and bridge the habitability gap.
  • Ethical issues such as trust, manipulation through affective computing, nudging, and liability require careful design and evaluation frameworks.
  • Developing computational models of interactional intelligence and incremental multimodal processing is crucial for natural conversational dynamics.
  • Standardized evaluation metrics and protocols are currently lacking in robotics compared to speech technology and need urgent development.
Principles it states
Beneficence: promoting well-being, preserving dignity, and sustaining the planetNon-maleficence: privacy, security and capability cautionAutonomy: the power to decideJustice: promoting prosperity and preserving solidarityTransparency and Explicability: enabling the other principles through intelligibility and accountability
Outline
· Executive Summary
· Overview of Stimulus Talks
· Show and Tell
· Individual Contributions from Participants
· Participants
Issued10 January 2020
Length24,805 words
FormatNot recorded
Scopeglobal

How it scores

Pending review

Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.

Accountability

3/5

Clear responsibility for AI outcomes

The question of liability arises for designers and trainers of virtual conversational agents.

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Fairness

3/5

Non-discrimination and equitable treatment

2. Learning bias: Some of the virtual agent models are trained from data selected by a “coach” (human agent in charge of selecting them). The agent can be discriminating, not fair if the data are badly chosen.

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Human Dignity

3/5

Respect for the inherent worth of persons

Some ethical principles have been proposed by the EU experts: Beneficence: promoting well-being, preserving dignity, and sustaining the planet

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Human Oversight

1/5

Humans remain in control of consequential decisions

Autonomy: the power to decide

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Privacy

2/5

Protection of personal data and private life

Non-maleficence: privacy, security and “capability caution”

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Transparency

3/5

Openness about how AI systems work and are used

Transparency and Explicability: enabling the other principles through intelligibility and accountability

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Enforcement

0/5

Consequences and compliance mechanisms with real force

screening score · 2026-10-23

Impact Assessment

4/5

Required assessment of an AI system's risks/impacts before or during use

Our aim is to study these interactions and relationships, in order to audit and measure the potential influence of affective systems on humans, and finally to go towards a conception of “ethical systems by design” and to propose evaluation measures.

3.1 Executive Summary / Ethics

screening score · 2026-10-23

Topics

Physical AI Governancecontext
human_robot_interaction spoken_language_processing virtual_agents ethics evaluation multimodality