Realbotix M-Series Humanoid Robot Teaches at New York High School

What happened?

The Salamanca City Central School District in New York has announced an educational pilot that integrates the Realbotix M-Series humanoid robot, along with the Optio tutoring platform, into its high school classrooms. The robot, designed to interact face-to-face with students, combines natural language processing, facial expressions, and real-time conversation. Its initial role will focus on the district's AI and robotics classes, where students can observe how it processes language and coordinates its digital responses with facial movements, demonstrating the concept of 'embodied AI.' On the other hand, Optio offers digital avatars trained on the district-approved curriculum, accessible during and outside school hours for individual tutoring and reinforcement in multiple languages.

According to the official press release from Onconetix —the company that has agreed to acquire Realbotix— dated June 24, 2025, the pilot starts with the district's AI and robotics courses and plans to expand to about 500 high school students by fall 2026, though that expansion explicitly depends on the results of the initial phase. The Salamanca district is located within the territory of the Seneca Nation, a detail that underscores the unique cultural context of this initiative.

Why is it important?

This pilot represents one of the first documented cases of a humanoid robot with advanced conversational capabilities used as a teaching assistant in a public high school classroom in the United States. The combination of a physical robot with a virtual tutoring platform could set a precedent for integrating social robotics into education. Additionally, the Salamanca district is recognized as a Woz ED STEM Pathway, a program founded by Steve Wozniak, co-founder of Apple, which provides a context of educational innovation and connection to the history of personal computing.

The project addresses key questions about the effectiveness of embodied AI in learning environments, acceptance by students and teachers, and the ethical and practical challenges of delegating educational functions to machines. Unlike previous experiences with robots in education —such as the use of NAO in European schools or Pepper in Japan— the Realbotix M-Series is specifically designed for realistic social interactions, with an animatronic face that can display emotions. This places it in a new category of 'social robots' that seek to generate emotional bonds, which in education could have both benefits (greater engagement) and risks (emotional dependence).

Historically, educational technology has gone through phases: from educational radio and television in the 20th century, through personal computers in the 80s, to virtual assistants like Siri or Alexa in the last decade. However, the incorporation of a physical robot with a human appearance marks a qualitative leap, as it implies a physical presence that can modify classroom dynamics in unpredictable ways. Experts in educational robotics, such as Professor Tony Belpaeme from Ghent University, have pointed out that robots can be effective for specific tasks like language teaching or basic math, but warn that their long-term impact on children's social and emotional development is still unclear.

What consequences will it have?

If the pilot is successful, it could accelerate the adoption of humanoid robots in schools across the country, especially in areas with teacher shortages or needs for educational personalization. According to data from the National Center for Education Statistics (NCES), in 2023, 45% of U.S. public schools reported difficulties filling teaching positions, a trend that could drive the search for technological alternatives. Additionally, the decreasing cost of humanoid robots —the M-Series has an estimated price of $30,000 to $50,000, according to market analysis— could make them viable for districts with tight budgets.

However, it also raises concerns about student data privacy, technological dependence, potential dehumanization of teaching, and impact on teaching jobs. The M-Series robot, being connected to the cloud for natural language processing, could collect sensitive student data, requiring compliance with laws such as FERPA and COPPA. On the other hand, the expansion to 500 students by fall 2026 will depend on the results of the initial phase, indicating a cautious approach. In the long term, this experiment could influence educational policies and the design of curricula that integrate AI as a pedagogical tool.

Compared to other pilot programs, such as the use of robots in schools in South Korea or Finland, the Salamanca case stands out for its bet on an advanced humanoid robot rather than simpler robots. In South Korea, the government invested over $10 million in 2019 to deploy robots in kindergartens, but results showed limited improvement in learning, although there was an increase in children's engagement. In Finland, a 2022 study with the NAO robot in language classes found that students interacted more, but teachers reported an additional technical workload. These precedents suggest that the success of the Salamanca pilot will depend not only on the technology but also on teacher training and curriculum integration.

What should readers know?

The M-Series robot will not replace teachers but will act as an assistant in specific activities, mainly in AI and robotics classes. Personalized tutoring will mostly be done through Optio's digital avatars, not the physical robot. The project is a limited pilot, not a mass deployment, and its future depends on rigorous evaluations. Interested parties should closely follow reports from the district and Realbotix for concrete results. Additionally, it is worth paying attention to reactions from the educational community and potential regulations on the use of AI in school settings.

It is also relevant to note that Realbotix is in the process of being acquired by Onconetix, a biotechnology company, which could create uncertainty about the continuity of technical support for the robot. On the other hand, the Optio platform is based on large language models (LLMs) trained on the district's curriculum, meaning its responses are limited to approved content, reducing the risk of incorrect information but also limiting its ability to adapt to unforeseen questions.

In conclusion, this pilot is a valuable experiment worth attention, but it should not be interpreted as an immediate solution to educational problems. Educational technology has a history of unfulfilled promises —from digital whiteboards to MOOCs— and only careful evaluation will determine whether humanoid robots can have a real impact on learning. Until then, the question that no one can yet answer is whether it works.