The real thing is that two important events related to AI (Artificial Intelligence) are happening all over the world, which are getting everyone’s attention. First, you may have seen the news that the US government has ordered Anthropic – a company that makes AI models – not to allow foreign organizations to use its new AI models. Anthropic is the same company that created the famous ‘Cloud’ AI model. Now the global race has begun in the field of AI – a technology that is gradually enabling computers to do tasks that humans used to do. AI models like ChatGPT and Cloud allow people to ask all kinds of questions and get answers; They can understand medical reports, generate a variety of reports, write computer code, compose emails with ease, and even create videos and movies. In fact, tasks that previously took months to complete can now be completed by AI in a matter of seconds.
And often the results are very good. As a result, companies at the forefront of AI are making progress around the world. Armies are now using AI in warfare, and AI-powered robots are taking over countless tasks for humans. AI companies are creating models that give users a significant competitive edge. For example, Anthropic created an AI model named ‘Mythos’, which was designed to detect vulnerabilities in computer programs. It caused a stir by showing its ability to break into banking systems and expose weaknesses in military networks. This gave rise to concern, as it was felt that such technology could cause massive destruction if it fell into the wrong hands.
Another AI model: Fable 5
In response, Anthropic developed another AI model, ‘Fable 5’, which includes safeguards that prevent it from breaking the security systems of banks, military networks and similar organizations. However, the US government imposed restrictions stating that models such as the Cloud 3.5 Sonnet (often called ‘Fable 5’ or ‘Mythos’ in this context) could not be made available outside the US. What this means for us is that Indian citizens, Indian companies, and even foreign employees working at Anthropic cannot use these models. Anthropic argued that it would be impossible to determine whether a user was American or not, so they blocked access to these models worldwide. It is now clear that countries will use AI more as a strategic tool, and American companies are currently at the forefront in creating AI models, followed by Chinese companies.
In a way, AI is becoming like crude oil or gold – whoever has it gets to rule the world. This means that India should also work on creating its own AI models. Although India has started this process, we are still far behind, because America and China have moved ahead. They are now using AI to make robots. Right now, factories in countries like India provide low-cost labor for goods that would be expensive to make in countries like the US. However, if they use AI-powered robots, the need for goods made in countries like India may disappear. This news points to a scary threat for us.
However, there is another part of the news which is more worrying. How are these AI models actually created? How are computers taught to think like humans? How are robots instructed to act like humans? Data related to human working methods is collected. And where does this data come from? From countries like India. As AI robots are being developed around the world, they require large amounts of real-world video data to train them.
Basically, AI models are trained using “first-person” video – footage that shows the robot’s perspective, as if seeing through its own eyes. So, how is it actually done? In India, companies pay employees, laborers and even domestic workers – but why? So that they can wear a small camera on their head while doing their work. Their way of working – whether it is sewing clothes, applying labels, packing goods in a factory, or cutting fruit, washing utensils or doing anything else at home – is continuously recorded. The camera mounted on his head captures the entire process on video. This footage then becomes training data for the AI. This is called ‘ego-centric data’.
In the US or Europe, obtaining such data is either against the law or very expensive. It is not possible to create AI models without data. So, they are collecting data. They are far away from us, yet when they create AI, they make laws preventing its distribution to other countries. The question is why should we let this happen? We are able to create our own AI. But if American companies want to keep the AI models created from it only with themselves, then why should we give them our data? If America is building walls around itself in the Trump era, then countries like India will have to answer to this. Ultimately, this issue could prove to be a bigger fight than the tariff issue. This is the real thing.
Training a robot using video
AI companies then use these videos to train robots to do the same job. And those who wear head cameras are paid a good amount for doing this work – Rs 250 per hour. Right now, people think they are making good money; After all, they’re doing the same thing they normally do, just with a camera pointed at their heads. It serves as data for robots, and similarly, data from other works is also being collected from countries like India to train AI models. For example, the process of writing an article or a human preparing a balance sheet – all this data is recorded on a computer and used to train an AI model.
