India’s power sector is high on policymakers' and investors’ lists. The draft National Electricity Policy is in advanced stages of formulation. Demand is expected to increase over the next decade, driven by manufacturing, data centres, and electrification of transport, among others. Indian and international companies are keen to capitalise on this opportunity.
But the power sector also offers opportunities beyond traditional investments in power generation and networks.
One such opportunity is investing in AI-based technologies and solutions. The Central Electricity Authority has constituted a committee to examine emerging AI applications, facilitate knowledge sharing, and recommend suitable initiatives and guidelines for AI adoption in the sector. But the broader question remains: What should the policy on AI use in the power sector be?
We believe the overarching priority should be people: Electricity consumers and workers. A pro-people approach to adopting AI will improve the lives of customers while simultaneously upgrading work of employees.
Consider some facts. Customers across much of India face hundreds of power interruptions each year, losing tens of hours of electricity and incurring significant economic losses. Voltage fluctuations and harmonics cause a lot of damage to appliances and equipment. Electrical accidents result in thousands of lives lost and injuries to many more.
In an AI enabled world, a consumer opens her electricity app and sees much more than a bill. It tells her why consumption has risen in that month. It suggests when to run the washing machine or water pump. It alerts her to an unusual spike in usage. If there is an outage, it tells her when power is likely to return.
{{/usCountry}}In an AI enabled world, a consumer opens her electricity app and sees much more than a bill. It tells her why consumption has risen in that month. It suggests when to run the washing machine or water pump. It alerts her to an unusual spike in usage. If there is an outage, it tells her when power is likely to return.
{{/usCountry}}A well-designed AI system can play a critical role in creating such an experience. It can forecast demand, detect faults before consumers experience them, and predict equipment failure. A transformer can be repaired before it fails. An overloaded feeder can be identified before it trips. An outage can be predicted and restoration planned faster.
This is the real promise of adopting AI in the electricity sector. Not technology for its own sake but technology for people. Fewer outages. Faster restoration. More accurate bills. Lower losses. Better integration of solar and batteries. And, over time, lower costs.
But there is another question. What happens to the people who keep the lights on?
A recent paper by Daron Acemoglu, David Autor and Simon Johnson, Building Pro-Worker Artificial Intelligence, provides direction. It argues that AI need not simply replace workers. It can also expand human capabilities, create new tasks and make workers more productive. India should apply this approach to adopting AI in the power sector.
Now consider the situation of discom employees when a transformer fails at 2 am. A neighbourhood goes dark. A lineman gets a call. He drives to the site, often without knowing exactly what has gone wrong. He checks the feeder and the transformer. He climbs the pole. He works in the dark, sometimes in the rain. He may need to call the shift engineer or the store manager.
But the same situation can play out as a very different scenario. Before he leaves, an AI system has already analysed the feeder. It has compared the transformer’s temperature, load, voltage and past performance. It tells him what likely failed and flags possible safety risks. It suggests what equipment he may need and generates a request for replacement.
The lineman still does the job. But he does it with better information. That is what pro-worker AI can look like. AI cannot climb a pole. It cannot repair a burnt transformer. It cannot make the judgement needed when conditions on the ground differ from what the data predicted. But it can process millions of data points and identify patterns that a human cannot. The worker becomes more capable and, more importantly, safer.
Future electricity jobs will look different. AI will create demand for people who understand both electricity and digital systems. India will need technicians who can maintain smart meters and sensors. Engineers who can interpret AI-generated fault predictions. Data analysts who understand power networks. Cybersecurity specialists who can protect an increasingly digital grid.
It will also need people who can question the machine. An AI system may flag a consumer as a possible electricity thief. A human must be able to challenge that conclusion and validate it based on site visits, personal knowledge and application of professional judgement.
An AI-driven future need not be human versus machine. It should be human plus machine. Technology alone will not make AI pro-worker or pro-consumer. Policy will. So, what considerations should drive the policy towards people-centric AI adoption?
First, AI technologies should be judged by outcomes. Did outages fall? Did restoration become faster? Did consumers receive better service?
Second, employees at various levels should be involved in selecting and implementing solutions. They know the challenges and how the system works.
Third, AI adoption must come with training. A lineman whose work changes should have a pathway to new, higher-skilled tasks.
Fourth, AI-based billing and theft detection should have human review and strong grievance mechanisms. An algorithm should never become an unquestionable authority over a customer.
Finally, AI-based solutions should augment and not just automate. A system that eliminates a hundred tasks is not automatically a success. A system that helps a hundred workers perform those tasks more safely and effectively creates much greater value.
India has a choice. Its electricity sector could become one of the country’s biggest laboratories for pro-people AI. The policy matrix should be: How many more people get reliable power? How much safer are the people who provide it? And how much more capable are the workers who keep the lights on?
The draft National Electricity Policy provides the perfect opportunity to prioritise investments and guide administrative action towards pro-worker and pro-consumer AI technologies in the power sector.
(The views expressed are personal)
This article is authored by Gaurav Bhatiani, senior fellow and visiting professor, ISB, and Vaibhav Chowdhary, director, Ashoka Centre for a People-centric Energy Transition.