The field of multimodal AI is moving towards a more deployment-centric approach, incorporating deployment constraints early in the development process to reduce the likelihood of undeployable solutions. This shift is driven by the need to integrate diverse types of data and improve understanding, prediction, and decision-making across various disciplines. Researchers are exploring the use of foundation models, such as large language models and multimodal models, to support software engineering activities, including coding and testing. Additionally, there is a growing interest in applying multimodal AI to real-world problems, such as intelligent agricultural decision-making and creating inclusive art environments. Noteworthy papers in this area include:

• A study proposing the Multimodal Agricultural Agent Architecture (MA3) for intelligent agricultural decision-making, which leverages cross-modal information fusion and task collaboration mechanisms.
• A paper presenting a knowledge base for arts and inclusion, utilizing the Dataverse data archival platform as a knowledge base management system enabling multimodal accessibility.
• A research roadmap for integrating foundation models into various phases of cyber-physical system software engineering, highlighting key research opportunities and challenges.