Pune: Who takes responsibility when an artificial intelligence (AI) bot makes a mistake, and whose job is at stake? This question is emerging as a key concern for technology leaders navigating AI adoption across industries.
The issue came up at a masterclass attended by 30 senior technology leaders from global capability centres (GCCs) in Pune, organised as part of a conference by consulting firm Zinnov and productivity analytics company ProHance. He cited a cybersecurity GCC that uses 10 bots to monitor threats, overseen by a manager bot.
Yet they are the ones held responsible if AI does something wrong,” said Prateek Gupta , an IT professional working at a GCC. Neha Sharma , a technology professional at a banking GCC, said managers wanted to move beyond pilots and deploy AI tools for consumers, but struggled to do so within existing regulatory frameworks. Meanwhile, some firms are adopting AI rapidly for fear of falling behind, often without fully assessing costs, said Nitin Panicker of ProHance, who conducted the masterclass. Some teams have exhausted their annual AI query budgets in a single month, he said, making it essential for companies to assess usage. Agentic bots inserted between two human tasks can go unused, resulting in underutilised AI, one participant said. Another technology professional said AI-generated suggestions could be useful, but the final outputs were not always trustworthy. Nitin Patil, who works at a renewable energy GCC, said his team initially struggled to use AI to identify suitable land parcels for solar or other renewable energy projects. Panicker said a large conglomerate reviews and replaces portions of AI-written code with human-written code to protect legacy systems from potential risks.
“The biggest barrier for AI adoption across teams is the fear of losing native knowledge. People also feel that they are losing the sense of ownership related to a task. In banking, rigid governance structures are another obstacle. However, the expense of such deployments can be substantial. Poor integration poses another challenge. Context also matters. Adding relevant data points improved the model’s output, making the solution acceptable. Some companies are also introducing human oversight into AI-generated code.
You Can Also Check: Gold Rate in Pune | Silver Rate in Pune | Bank Holidays in Pune | Public Holidays in Pune | Petrol Price in Pune | Diesel Price in Pune | CNG Price in Pune | LPG Price in Pune Stay updated with the latest Pune news. Download the TOI App.