Artificial Social Intelligence Workshop

ICCV 2025

Full-Day Workshop

Date and Time TBD

Room TBD, Hawaii Convention Center

Hybrid Option Available

Social Intelligence in AI Systems

Humans use social intelligence to interpret and navigate multimodal interactions with other agents in our shared world. As artificial intelligence (AI) systems become pervasive in human life, it is crucial to improve the social intelligence of these systems for them to seamlessly work with and around humans. There has been a growing interest across computing communities to build competencies for core social intelligence abilities such as social perception (e.g., perceiving gesture) and social reasoning (e.g., inferring intent) in AI systems. These systems with artificial social intelligence could enable richer human machine interactions to support human well-being in homes, hospitals, manufacturing, and other settings.

Research priorities and modeling frameworks to build social intelligence competencies in AI systems can vary across computing communities (and have varied in prior decades). What are core technical challenges and opportunities for cross-field collaboration to advance the science of social intelligence and socially-intelligent AI? A particular focus of the workshop keynotes and discussions will be algorithms for reasoning, multimodality, and embodiment in socially-intelligent AI systems. Our ICCV 2025 workshop welcomes anyone interested in artificial social intelligence to join us to discuss these topics and more!

Call for Papers and Extended Abstracts

To facilitate interdisciplinary research in artificial social intelligence, we invite researchers from computing and robotics (Computer Vision, NLP, ML, HCI, HRI), psychology, neuroscience, and sociology to join our workshop and discuss underlying technical challenges, goals, and ethics of building artificial social intelligence.

Our workshop will accept submissions via OpenReview to 2 tracks: Papers (archival) and Extended Abstracts (non-archival).

Papers are 4-8 pages (excluding references) and will be published in the ICCV workshop proceedings. Papers will be considered for oral or poster presentation at the workshop.
Extended Abstracts are up to 2 pages (excluding references) and will not be published in the proceedings, and will be presented as posters. Extended abstracts may be ongoing work, recently published papers at other venues, or papers published at the main ICCV conference.

Topics

Topics of interest for papers and extended abstracts include, but are not limited to, the following:

Social reasoning systems
Affective computing (e.g. predicting emotion, valence)
Social perception and social signal analysis (e.g. gaze recognition)
Generating social signals in artificial agents (e.g., gesture generation)
Social agent frameworks
Social robots and socially-intelligent human-robot interaction
Datasets, benchmarks, and community resources
Ethical considerations for Social AI
Applications of artificial social intelligence

Submission Instructions

Submissions will follow the ICCV paper template and guidelines and must be anonymized. Submissions for the paper track may include an optional appendix after references.

Papers Track

Submission Deadline: June 27th, 2025

Notification: July 11th, 2025

Submission Portal: OpenReview Archival Workshop Submissions

Extended Abstracts Track

Submission Deadline: August 1st, 2025

Notification: August 15th, 2025

Submission Portal: Openreview