Shalu is a home-made humanoid robot built by Dinesh Kunwar Patel, a computer science teacher at a Kendriya Vidyalaya (a central government school) in Mumbai. He made it from waste materials such as plastic, cardboard, wood, aluminium and old newspapers, spent about three years on it and, by his account, around Rs 50,000, which was roughly 675 US dollars at the time. Patel says Shalu can talk in 47 languages: nine Indian languages and 38 foreign ones, the figure that made the headlines.
The robot is not a lab-grade machine, and Patel never claimed it was. What makes it interesting is what one teacher managed with free software, scrap materials and persistence, and how the project ended up in classrooms.
Who built Shalu, and why
Patel has named two inspirations. One is Enthiran, the 2010 Tamil science-fiction film starring Rajinikanth, which was released in Hindi as Robot. That is why early reports described it as a Bollywood film, although it was made in the Tamil industry. The other is Sophia, the humanoid robot from Hanson Robotics that became a media star from 2016 onwards.
He wanted to show that a useful robot did not need an expensive lab. Shalu's body is shaped from discarded materials, and Patel wrote its software himself, using two free, open-source toolkits: TensorFlow, a machine-learning library first released by Google, and the Natural Language Toolkit (NLTK), a long-standing Python library for working with text and language. The project is dedicated to the government's Beti Bachao Beti Padhao campaign for girls' education.
The first version was shown publicly on 23 November 2020 in a broadcast on DD News, India's public news channel. Patel has said a newer version is in development.
What the robot can actually do
According to Patel and the descriptions published since, Shalu can:
- greet people, shake hands and hold simple scripted conversations;
- recognise and remember faces and identify common objects;
- pick up basic emotions such as happiness or anger and respond with expressions of its own;
- answer general-knowledge questions, run quizzes and solve simple maths;
- read the news, recite poems and recipes, give weather forecasts and read horoscopes;
- present lessons from slideshows prepared by a teacher, then ask and assess questions.
Patel also suggested it could work as a receptionist in an office, a companion for older people or a learning partner for young children. A computer science professor quoted in the early coverage, Supratik Chakravarty, was enthusiastic, saying such a robot "can be used in the field of education, entertainment and several other fields too."
The video below shows Shalu answering questions and talking with Patel.
The 47-language claim, explained
Headlines said Shalu speaks 38 languages, which counts only the foreign ones; the total Patel gives is 47. That figure needs a little context.
Modern speech-recognition and translation tools already handle dozens of languages, and a hobby project can connect to them or to similar language models. So it is entirely plausible that Shalu can produce and respond to speech in many languages. But being able to speak a language and understanding it the way a fluent person does are different things. Much of what the robot says follows prepared scripts, and a short demonstration shows chosen examples rather than performance across every language and situation. Independent testing of the language range has not been published, so the number is best read as the builder's own description.
None of this takes away from the achievement. Combining speech, face recognition, simple emotion detection and a moving physical body on a teacher's budget is a solid piece of engineering and a good demonstration of what open-source tools make possible.
Awards and a place in the syllabus
On 25 March 2021, Patel and Shalu received the Jagran Josh Education Award for Innovation in Science and Technology, presented by the then education minister, Ramesh Pokhriyal.
Since then the robot has moved from news story to teaching aid. Shalu has been included in the CBSE Class 9 artificial intelligence syllabus and in the Class 6 computer science syllabus of the Kendriya Vidyalaya Sangathan, the network of central government schools. At Patel's own school it is used in lessons for classes 6 to 11. For students, a robot built from cardboard and scrap by their own teacher is a powerful message: AI is not only something made by distant tech giants.
Why the online reaction was mixed
Not everyone was charmed. When videos spread on social media, many commenters were blunt. Some argued that Shalu was essentially a voice assistant, similar to Alexa or Google Assistant, placed inside a hand-made head, and that a simple speaker would have done the same job. Others said its face looked eerie.
Both reactions have a point, and both miss something. It is true that much of the conversation ability in projects like this comes from existing software. But building the face tracking, emotion responses, movement and lesson tools around it is the hard and creative part. The eerie feeling has a name too: the "uncanny valley", an idea proposed by the Japanese roboticist Masahiro Mori in 1970. It describes how a robot that looks almost, but not quite, human can feel more unsettling than one that looks clearly mechanical. Even well-funded robots like Sophia draw the same comments.
What projects like Shalu teach us
- Open-source tools have lowered the barrier. Libraries like TensorFlow and NLTK are free, well documented and used in universities and companies alike.
- Classroom robots work best as helpers. A robot can run a quiz or present a lesson, but the teacher still plans, explains and cares for students.
- Read capability claims carefully. A number like "47 languages" tells you what a system can attempt, not how well it understands. Ask what was tested and how.
- Making something is a lesson in itself. For students, watching a teacher build, fail, fix and improve a robot over three years teaches more about engineering than any textbook.
Sources: Wikipedia: Shalu Robot; online reactions via 9GAG; video from YouTube.
COMMENTS