Several key leaders from top tech companies gathered for an insightful fireside chat at the Invest Karnataka 2025 event in Bengaluru to discuss the transformative potential and implications of AI in India.
One of the standout participants was Manish Gupta, a senior director at Google DeepMind, who discussed the importance of building AI in India while striking a balance between fostering innovation and establishing guardrails to enable responsible growth.
Over the last few days, India has witnessed significant strides towards building a foundational AI model. As the country prepares to embark on this journey, a pressing question arises: What impact will it have on jobs?
Gupta said that the technological industry is responsible for building AI that can augment human capabilities rather than substitute for them.
“Each and every person on the planet will have at their disposal an amazingly powerful, universal assistant, as well as specialised assistants that will help them learn any skill, lead healthier lives, take care of shopping or other chores and so on.” Gupta highlighted that it’s important not to be afraid of using such AI assistants.
Gupta was also joined by Ranganath Sadasiva, CTO of Hewlett Packard Enterprise India, and Caleb Munigety, head of Bosch Enterprise India AI, who unanimously agreed to the sentiment put forward.
According to Munigety, AI will not terminate jobs but transform them. “In the case of the Industrial Revolution, what did it do? It made people walk from farms to factories. So, a transformation of the job profile.”
In addition to industry experts, Prime Minister Narendra Modi, echoing a similar sentiment, said people shouldn’t fear the loss of jobs with the advent of AI.
Is AI a Threat to Millions of Developers in India?
However, opinions aside, it is only fair to trust the numbers. As the famous statistician W Edwards Deming once said, “In God we trust; all others must bring data.”
An increased adoption and development of AI in India would raise concerns about how it may disrupt the lives of millions of software developers.
For instance, Anthropic recently released a report that outlined how people are using AI in their jobs. The study found that AI is primarily used for software and mathematical tasks. Conversely, tasks involving physical manipulation of the environment minimally use AI.
However, the report revealed that 57% of use cases involved leveraging AI to augment human capabilities, while 43% suggested automation.
Even if the usage of AI in software development sparks fear, the reality is that it only works best as a companion. Take the autonomous software developer tool, Devin, for example.
“We’ve really enjoyed the chance to experiment with Devin, but we find it hard to agree that it’s at the level of a junior software engineer,” read a blog post written by the co-founders of RunLLM, an AI-enabled technical support platform. They mentioned that it works in automating fixed, well-defined tasks, but results vary dramatically beyond the scope.
“As for where things stand today, there’s a long way to go for any software engineer to worry about losing their jobs,” they added.
On the other hand, platforms that augment software development have observed tremendous growth. Cursor, one such platform, was the fastest SaaS tool to reach $100 million in revenue.
Will Brown, an AI researcher at Morgan Stanley, said that, compared to Devin, he’d much rather prefer Cursor’s workflow, which requires a human programmer to prompt the tool. He said the experience is “way more hands-on”, and much easier to provide suggestions.
Human oversight is indeed necessary when using AI to code. For instance, when Devin was asked to make changes to the CSS code, the changes brought in a few unnecessary components.
Not being able to understand such changes and ending up with a pile of AI-written code is a problem. “AI always gives you an answer, and the answer is not just wrong, but also very hard to detect what is wrong about it. It just invents stuff,” said Mark Essien, CEO of Hotels.ng.
Besides human oversight, fundamental skills are crucial in dealing with AI-generated code.
Recently, a user on Reddit revealed that his coding project had grown so large that Claude could not understand it.
The user built a Python project using Claude and Cursor, but he ended up with a “disorganised” codebase of over thirty files, which the user suspects “might have duplicates”. They had trouble getting Claude to fix the most basic errors and bugs. “It doesn’t even recognise the main issue and just ends up deleting random lines or breaking everything completely,” the user further said.
However, AI alone isn’t the problem in this situation. The user revealed having “zero knowledge” of Python and called the project’s progress so far a “miracle.”
Similarly, AIM, through its reportage, found that AI is augmenting jobs rather than replacing them across various industries. From assisting designers in refining their concepts to helping journalists generate story ideas or automating aspects of filmmaking, AI seems to be evolving as a powerful tool for collaboration rather than competition.
