Embodied AI in Humanoid Robots: Progress, Social Interaction Dynamics, and Barriers to Scaling an Emerging Industry

Humanoid robots are rapidly evolving from pre-programmed mechanical systems into socially aware, AI-driven agents capable of operating in human environments. Early generations of humanoids relied on scripted dialog, rule-based conversation trees, and domain-specific speech modules, systems that produced predictable but limited human–robot interactions. These robots could speak, but they could not listen, adapt, or reason.
Recent advances in large language models, embodied AI agents, multi-modal perception, and cloud-scale training have transformed this landscape. Modern humanoid robots can now engage in naturalistic, context-aware conversation, interpret visual and auditory cues, plan actions, learn from demonstrations, and execute complex tasks with higher autonomy. These capabilities are reshaping the social dynamics between humans and robots, enabling more fluid collaboration, trust formation, and shared problem-solving.
Yet despite this progress, the humanoid robotics industry faces significant challenges in scaling hardware, safety, battery life, reliability, and economically viable use-cases. This talk will provide a balanced overview of recent technical breakthroughs, the realities of implementation, and the engineering barriers that must be addressed for humanoid robots to move from controlled pilots to widespread deployment.
Following the talk, all attendees are invited to dinner to be held at a nearby restaurant.
Speaker(s): Sudhir Shenoy,
Agenda:
6:00 – 6:15 PM – Introductions
6:15 – 7:30 PM – Embodied AI in Humanoid Robots
7:30 – 8:00 PM – Conclusion and Q&A
Room: 401-A Conference Room, Bldg: Martin Luther King Jr. Memorial Library, 901 G St. NW, Washington, District of Columbia, United States, 20001, Virtual: https://events.vtools.ieee.org/m/519535