Independent research & editorial platform
Somatic AI research for embodied movement and human–AI co-creation.
Founded by Hanna Zhu, Somatic-AI Lab connects somatic practice, movement computing and generative systems through source-linked research, field synthesis and practitioner guidance.
This week · Pulse
Mon / Tue / Wed drops — news scan, community field notes, and a plain-language explainer. Updates weekly (Mon–Wed). Public categories: news-digest, community-news, popular-explainer.
Motion AI this week: control moves from prompts into representations
Four preprints released from 25–27 September 2026 show motion control moving deeper into model design. Rather than asking a prompt to carry every instruction, the new systems place style intensity, partial observations, social affect, or transformation history inside the representation itself.
Dance technology this week: rhythm, groups, and interaction states
Three preprints released on 29 September 2026 treat dance and social movement as relationships rather than isolated poses. One studies minimal rhythm feedback for beginners, one generates beat-aligned 3D dance, and one models the changing state of a whole group alongside each participant.
Why can dance feedback feel helpful before your score improves?
Feedback can improve a beginner’s ability to recover, continue, and feel oriented before it produces a measurable improvement in timing. That is not evidence that measurement is useless or that feelings equal skill; it means learning has more than one time scale and more than one outcome.
This month · Long-form
October frontier report: motion models are being judged by what actions they preserve
The central conclusion for September’s final research window is that motion representation is becoming action-relative. From 23–30 September 2026, new preprints moved beyond asking whether a latent code reconstructs movement or predicts a plausible next frame; they asked whether it preserves a contr…
Prediction is not yet situated action: the philosophical limit of world-model imagination
A world model can predict future observations without yet understanding which differences matter for action. Recent audits show that prediction accuracy, exploration gains, and planning success can separate. Read through situated-action and enactive theory, this is not a surprising failure of scale;…
From gesture to event: what structured motion representations make visible
Movement is often introduced as a shape in space. A hand rises, a torso turns, a person pauses. Recent motion research complicates that picture: the important unit may be an event with a cause, a contact, a social meaning, and a degree of uncertainty. This month’s papers provide a useful way to thin…
Innovation brief: make contact a first-class motion command
Most language-steered motion systems describe where a limb should go. Fewer describe what happens when it meets something. Atomic Motion offers a compact control vocabulary—signed translation, rotation, and hold atoms—while Beyond Gestures shows that a wrist pressure array can recover hand pose and …
Build an evidence ledger for a movement-reactive AI system
This guide is for creative technologists building movement-to-sound, movement-to-image, or embodied robot prototypes. It provides a small Python 3.11+ logging pattern that records whether every control value was observed, inferred, generated, or actively requested, so a rehearsal failure can be trac…
Lian Loke: designing technology from the lived body outward
This is a source-verified profile, not a direct interview. Lian Loke is an interdisciplinary artist and interaction-design researcher whose work places lived movement, somatic literacy, and choreographic method at the centre of human–computer and human–robot interaction. Her importance to somatic AI…