The field of human-AI interaction is moving towards designing more intuitive and transparent interfaces that can foster sustainable behaviors and ensure secure interactions. Researchers are exploring the potential of anthropomorphic design in conversational agents to encourage environmentally conscious behaviors, with results indicating that such designs can improve user engagement and promote eco-friendly habits. Furthermore, the development of domain-specific languages and compliance mechanisms is gaining attention as a means to regulate AI interactions with the internet and prevent security vulnerabilities. The characterization of unintended consequences in human-GUI agent collaboration is also a key area of focus, with studies highlighting the need for more robust and user-centric designs that can balance automation with human oversight. Noteworthy papers in this area include:

• The introduction of ai.txt, a novel domain-specific language designed to regulate AI interactions with the internet, which enables precise element-level regulations and incorporates natural language instructions.
• The development of a generative agent-based simulation platform for human-assistant interaction research, which can simulate realistic human behavior and social dynamics, providing a scalable and cost-effective approach for studying assistant agent design.