The present paper aims at firstly, providing a multifaceted exploration of trends and challenges for ethical and trustworthy Multimodal Language Model development by offering a perspective on the recent developments around these models. Secondly, we explore Artificial Intelligence systems and their paradigms from an Industry 5.0 perspective across various application fields, such as Health Support Systems, Intellectual Property, Job Displacement, or Safety. Furthermore, we discuss relevant new paradigms, future perspectives, and mitigation strategies for some of their issues with an emphasis on multimodal systems and draw some conclusions.