Telangana has started training its own workforce to use artificial intelligence in day-to-day administration. Deputy Chief Minister Mallu Bhatti Vikramarka on 28 October inaugurated AI Saksham, a capacity-building programme meant to help government employees apply AI tools to speed up office work and improve the delivery of public services.
The programme was launched at the Dr Marri Chenna Reddy Human Resource Development Institute (MCR HRD) in Hyderabad. It has been organised by TG Math with support from the state’s Planning Department. Planning Secretary Gaurav Uppal, TRANSCO CMD Krishna Bhaskar and MCR HRD Director General Shanti Kumari were among those present at the launch.
Measuring AI by what citizens feel
Bhatti Vikramarka framed the initiative around a simple test: whether people still have to keep returning to government offices. Farmers, pensioners and ordinary citizens, he said, should be able to get their work done without repeated visits, and every hour officials save through technology should show up as faster and better service for the public. Timely resolution of grievances, he added, would be the real measure of whether AI is working in government.
Where the state sees AI helping
The Deputy Chief Minister outlined several areas where AI could support administration:
- Public expenditure: helping the finance department check whether funds are reaching intended beneficiaries, find the reasons behind spending delays and spot gaps in project implementation
- Welfare delivery: identifying shortfalls in how schemes reach eligible people
- Power sector: forecasting electricity demand, detecting equipment faults and improving maintenance so that farmers, small industries and hospitals get uninterrupted supply
- Office work: saving time in preparing government orders and reports, and making information available in Telugu
Guardrails before scale
Much of the message to officials was about caution. Bhatti Vikramarka said AI-generated alerts should not be treated as automatic proof of wrongdoing, and that officials must verify names, amounts, eligibility conditions and technical details before using AI output in official records or decisions. Employees, he said, should not enter official or personal data into unauthorised AI tools, and must understand what data can be used and what needs permission.
He called for mandatory human oversight, clear limits on the powers given to AI systems and clearly assigned responsibility if something goes wrong. Citizens, he said, must be able to reach an accountable official to seek redress against an adverse decision. Systems should be tested in real-world conditions and meet safety standards before being deployed.
On rollout, he advised departments to begin with small pilot projects and expand only after evaluating accuracy, cost, performance and public benefit, with the government keeping continuous oversight and control over the systems.
Why it matters
States are moving quickly to put AI into government workflows, and the bottleneck is increasingly the people using the tools rather than the tools themselves. By pairing employee skilling with explicit rules on verification, data handling and accountability, Telangana’s AI Saksham attempts to build that capacity inside the system first, before AI is used more widely in decisions that affect citizens.
Photo: Ravi Studios / Wikimedia Commons, CC0