This is like a dream coming true. Since my childhood I was fascinated by possiblity of humanoid robots.

Growing up watching american science fiction movies and shows, I always wondered will this be a reality in my life span. Well I was very optimistic about it and guess what I was not wrong.

Unitree Robotics has introduced a humanoid robot for a very achievable price – only $16000 and it is called Unitree G1.

Unitree G1 is around 4 feet and 2 inch tall and 35 kg heavy. It can walk on 2 legs at a speed of 5km/h, holds items with its 3 fingers, fold itself and sit on the chair.

Well this blog is not about the technical specifications of G1. You can read about it on their website also. This is about the moment in time in the human history when the mankind is about to take that giant leap. This kind of moment happened in the past when wright brothers flew the aeroplan, Apollo landed on the moon, Smartphones were launched and chatGPT was released.

The robots are not new. The size, weight and the price of G1 is the new. Unitree is making it possible for a tech enthusiast, a hobbyist, an early adopter, to be able to purchase a humanoid robot and bring a fresh perspective to the learning and practical use case of such machine.

It is very much like democratizing the robot training or in other words, crowd sourcing the product development. It is very different from what Boston dynamics has done with Atlas. Atlas is an amazing product and the new Atlas is even smaller and lighter too but is it available at this price to beta testers and users around the world? no.

I contacted Unitree for purchasing G1 and I received a response from their team that the interest in G1 is skyrocketed around the world and there is at least 6 months waiting period for any delivery.

I also asked them many questions to which I have not received any reply yet – for example:

1. Is it possible to train G1?

This is the most important question. If it is not possible to train a humanoid robot then its capacity, capability and usefulness, all become very limited. In that case it is just an expensive toy which can dance, sing and make funny sounds. The real use of a humanoid robot, even when it is in very early stage, is the possiblity of training him. The AI module of the such humanoid robot must be capable of learning

• The surrounding, layout of the house, the living room, location of furniture, tv, glass doors etc.
• To walk around the house without bumping into people, dog, chairs, shoes or anything else lying on the floor.
• To walk on uneven surface or at least climb 2-3 steps to go into the garden or come back.
• To respond to voice commands and understand the context of those commands like Come here, go there, sit down, hold this, pick it up, put it down.

The other questions which I am not able to find answers to are –

1. if G1 is capable of learning new tasks, will the learning be transferable to other G1s?
2. Is G1 capable of connecting to internet ? Is it programmable upto an extent that it can be integrated with chatGPT or Alexa ?
3. What is the privacy settings on the G1? Can it take pictures and send it back to Unitree servers ?
4. Does Unitree allow G1 to be connected with a laptop over an API ? Does it provide an interface using which it can be troubleshooted or software patches can be applied ?

It will also be interesting to see what guardrails have Unitree put around this humanoid robot. But it is not possible to get the answers just from their website and unless people get their hands on G1 we will never get to know the real risks and potential threats of humanoid robots.

It will very interesting to see the public policies and governance around such technology products in the society. So far such robots are available only for industrial applications e.g. Boston Dynamics’ SPOT is available for purchase for industrial applications and there is product guide, training, maintenance manual etc available to customers. But if Unitree or any other company decides to simply sell a 35 kg steel product with hands and legs to public it can end in a disaster. We are very well aware what some stupid people do with fast cars. The moment people get their hands on such machine they will make it do stupid and dangerous things. The Humanoid robot poses life threatning risks is not handled properly, such as

1. It is 4 feet tall, it can lose balance and fall on a pet animal.
2. It can bump into a child or an adult, causing injury.
3. It can crash into a glass door or a wall mounted TV and cause damage.
4. It can venture on a road and cause a car accident.

Because of these risks it will be very crucial for goverments and authorities to define laws for Humanoid robots. It will be very interesting to see the development in the insurance space. Will the insurance companies insure the humanoid robots just the way they insure the cars ? If G1 causes injury or death to a human or a pet, who will be held responsible, the robot, the owner, the handler, the manufacturer or the victim ?

Watch this space for more stories about humanoid robots !! and please leave me a feedback in the comment section.