The last few years have witnessed a proliferation of AI models and their rapid adoption across the world. Private individuals, firms, and governments are experimenting with various AI tools. In India, well known AI models are being rolled out either free or at minimal cost as tech firms scramble to acquire users and data to train their models. Many have partnered with telecom operators, giving them instant access to hundreds of millions of customers. Google, for instance, is offering its Gemini Pro free to select Jio users for 18 months. Similarly, Airtel subscribers can avail a free Perplexity Pro subscription. But the growing usage of these models comes with a set of concerns.
As per a report in this paper, questions are now being raised within the government on not only the data being uploaded but also the information being sought, and the inferences that can be drawn by the models. For instance, what are the implications of the data uploaded by millions, though anonymised, for the global firms and India — after all, data is the oil for the AI engine. Further, can the queries raised and the tasks being sought — by individuals such as bureaucrats, advisers, scientists, corporate leaders and others — be used for other purposes such as gauging their priorities, or determining their strengths or weaknesses? Worries over the implications of confidential information being uploaded have already led the finance ministry to reportedly ask its employees not to use tools like ChatGPT and DeepSeek on office devices.
These concerns come at a time of rising tech nationalism around the world. China, which has sought to build its own tech platforms, is already pushing its firms to adopt indigenous chips in AI data centres. In India, Union IT minister Ashwini Vaishnaw recently announced shifting to the Chennai-based Zoho’s office suite. The Union government has also selected several firms to develop indigenous large language and small language models under its AI mission. While, over time, this could increase the choices available in the domestic market — foreign platforms currently dominate most product segments — users are likely to gravitate towards indigenous platforms only if they are more innovative and offer superior services/experiences. Either way, the guardrails need to be put in place, and norms of basic hygiene must be adhered to.
