· Research & technology· 日本語で読む
Nikkei XTech reports Panasonic Holdings’ development of a model that predicts screen changes
Nikkei XTech reported on Nov. 13, 2026, that Panasonic Holdings and others had developed a world model. The report focused on technology that predicts changes on screens of mobile devices and other equipment. Related articles from this newspaper have covered separate research on tactile sensing and 3D recognition.
What happened
Nikkei XTech reported on Nov. 13, 2026, that Panasonic Holdings and others had developed a world model that predicts screen changes.
Analysis
The report on a world model that predicts screen changes indicates that research may address not only robots’ physical actions but also the information processing involved in tasks. However, the information provided does not confirm whether this model is connected to robot control or shares a common foundation with existing research. It is too early to evaluate screen-operation technology and technologies for handling objects or contact as a single system.
In Panasonic Holdings develops physical AI for robots using tactile sensing, with a 65% task success rate, we covered contact-state recognition and training using data from remote-operation video and tactile sensors. The screen-prediction technology reported this time addresses different types of information. Considered together, the reports suggest that when automating real-world tasks, interaction with the physical environment and understanding device states need to be evaluated as separate technical challenges. There is no information indicating that the two research efforts share data or models.
Also, Panasonic Holdings unveils “Proxy3D” to enhance VLMs’ 3D capabilities at CVPR 2026 covered research on spatial recognition in the real world. Screens, tactile sensing and 3D space appear as distinct topics in related articles, but no development plan integrating them has been announced. The breadth of research topics should be distinguished from the existence of a shared implementation and evaluation framework.
What remains to be confirmed is how predictions are evaluated, the range of devices and tasks covered, performance in real-world environments, and whether the technology is linked to physical tasks. Without this information, it is not possible to assess how far the research can be applied to on-site workflows. See the original source for details.
Sources
Translated from the Japanese original: /news/nikkei-kurosutekku-panasonikku-hd-no-gamen-henka-yosoku-mode-1cdd75. The Japanese version is authoritative if they differ.