Why can't humanoid robots just start working right away?
Humanoid robots are not like factory machines that repeat the same motion. They must operate in dynamic, unpredictable human environments — offices, hospitals, schools — where they need to recognize objects, understand speech, and interact naturally. A 2025 review of humanoid robot development notes that progress from taking minutes to make a single step in the 1970s to running and climbing today came from advances in AI that let robots learn from experience [2]. But that learning doesn't happen automatically; it requires structured training, much like a human apprentice.
For example, the humanoid robot Pepper was tested in a classroom setting to see how well it could capture speech. Researchers found that its speech recognition accuracy dropped sharply with distance and was affected by the speaker's age and gender [3]. To fix this, they integrated a separate speech-to-text tool (Whisper) and tested it with eight participants speaking at different distances. This kind of targeted training — identifying weaknesses and patching them — is exactly what 'school' provides before a robot can be trusted in a real workplace.
What do robots actually learn in their 'school'?
Robots learn two main things: technical skills (like recognizing speech or objects) and social skills (like following classroom norms or expressing empathy). A two-year study with teachers and children using a humanoid robot for math games found that teachers viewed the robot both as a didactic tool and as a social actor — and they switched between these views depending on the situation [1]. This dual role means robots must be trained to handle social expectations, like taking turns or responding to a child's frustration, not just solving equations.
In medical education, a humanoid robot named Ameca is being designed to act as a patient for students to practice diagnoses. The robot can mimic human facial expressions and conversations, providing a safe training environment where students can interact physically with a simulation patient without any risk to real people [5]. This shows that robots themselves need to be 'trained' in realistic scenarios — their school — before they can be used effectively in high-stakes jobs like healthcare.
What happens if a robot skips training?
Skipping training can lead to failures that range from annoying to dangerous. A 2021 framework for using humanoid robots in schools emphasizes that ethical considerations — like privacy, bias, and safety — must be addressed before robots enter classrooms [4]. Without proper training, a robot might misinterpret a student's accent, fail to hear a command in a noisy room, or violate social norms, making it useless or even harmful.
The same framework notes that education must prepare students for a future with robots, but it also implies that robots must be prepared for humans [4]. The 2025 review confirms that while over 150 companies worldwide are now developing humanoids, real-world applications are still limited mostly to automotive and logistics sectors [2]. This caution reflects the fact that robots that skip 'school' — rigorous testing and iterative improvement — are not yet reliable enough for broader use.
About These Sources
This answer is built on 5 peer-reviewed studies — published from 2021 to 2025, 1 from 2024 or later, 1 in Q1 journals, collectively cited 81 times — selected as the most relevant from 5 studies that passed quality screening, drawn from 47 papers retrieved from a database of over 500 million.
Sources used in this answer
The dual role of humanoid robots in education: As didactic tools and social actors
In a two-year co-design study with teachers and children, teachers viewed a humanoid robot both as a social actor and a didactic tool, switching perspectives depending on the context, which highlights the need for robots to learn social norms alongside academic tasks.
Humanoid robots: from the laboratory to the workplace
A 2025 review of humanoid robot history shows that progress from slow walking in the 1970s to running and acrobatics today came from AI advances, but real-world applications are still mostly in automotive and logistics, with over 150 companies developing humanoids.
The Synergy between a Humanoid Robot and Whisper: Bridging a Gap in Education
An experiment with 8 participants found that Pepper's speech recognition accuracy varied significantly with distance, age, and gender, and integrating a separate speech-to-text tool (Whisper) was needed to improve transcription for classroom use.
A Framework for Using Humanoid Robots in the School Learning Environment
A multidisciplinary framework for using humanoid robots in schools integrates technological, pedagogical, efficacy, and ethical perspectives, emphasizing that ethical issues like privacy and bias must be addressed before deployment.
Conception of a Humanoid-Robot-Patient in Education to Train and Practice
A concept paper describes using the humanoid robot Ameca as a simulated patient for medical training, where students can practice diagnoses safely without exposing real patients, demonstrating the need for robots to be trained in realistic scenarios.