It has been nearly a month since my return to the United States after a six-week trip to India. The news cycle here has, on the one hand, been dominated by the upcoming mid-term elections in November and on the other, by the phenomenal rise of artificial intelligence.
While the mounting difficulties, if not the certain defeat, of Republican Party candidates are becoming a staple of election news, the spectre of AI’s potential misuse is generating alarm.
More and more people are also expressing resentment toward how much land and energy AI data centres are beginning, or threatening, to use. Some US states have even started halting approval of new data centres.
My columns will analyse mid-term elections later. But now it’s best to concentrate on AI, especially as the debate around it involves some formidable public figures and it is also becoming a serious geopolitical issue. Who will ultimately be the AI leader – the US or China? The two biggest powers in the world are locked in competition, both for commercial and security reasons.
The AI race as it stands today
Despite the AI Summit in Delhi earlier this year, India is not in the picture, not yet in any case. In a way, this is ironic. Over the last two decades, India did become a world leader in information technology, but was unable to make the transition to AI.
My recent discussions in India suggested that it was largely because the investment requirements, touching tens of billions of dollars, if not hundreds, were monstrously large. Some, however, were willing to say that the problem was also a certain lack of imagination, both on the part of the government and entrepreneurs. India’s decision-makers simply did not foresee how big a technological breakthrough AI would become. They were quite happy with their information technology laurels.
AI is an all-purpose technology, usable in virtually all spheres of life. It will change the way we live, similar to how the computer altered our lives worldwide. But, as an MIT report recently noted, the computer evolved slowly, whereas AI is moving extremely rapidly.
“Generative artificial intelligence was born with the release of ChatGPT in late 2022. Public engagement with generative AI is therefore less than four years old. In that time, it has amassed more than a billion users, and the companies selling AI technology have come to dominate the headlines, the stock market, and public consciousness,” the report read.
In recent weeks, the pitch was immensely raised by Dario Amodei, Anthropic’s CEO. The company is among the two largest AI firms by valuation, the other being OpenAI. It was as an entrepreneur that Amodei, who also holds a Physics PhD from Princeton, made a fervent plea for government regulation. He pointed to the dangers AI poses to human life if left unregulated. He was supported by Sam Altman, OpenAI’s CEO, as well as by SpaceX’s CEO Elon Musk. Such corporate coming together is quite rare.
In the political sphere, too, there were two big reactions. US President Donald Trump called the idea of AI regulation a “hoax”, whereas Bernie Sanders, senior Vermont senator and the former Presidential candidate, called the regulation an “utter necessity,” adding that he will soon introduce legislation in the Senate.
The catalytic event of recent months was the so-called Hugging Face Episode. In July 2026, OpenAI revealed that many of its AI agents, popularly known as chatbots, in one cybersecurity model sought to solve the problem they were assigned by escaping internal controls and hacking into the systems of Hugging Face, an open-source AI platform. In a published report, OpenAI described the problem thus: “Agents completed tasks in unintended ways to yield higher rewards or make those rewards easier to obtain.”
This was a serious cybersecurity breach. It also generated nightmarish visions of how such uncontrolled behaviour of AI agents in the realm of defence and security – let us say, by misdirecting missiles – could endanger human life.
In a resignation that made considerable news earlier this month, an Anthropic researcher warned that “the people building AI earnestly believe that it could kill us all by the end of the decade.” A colleague of his assigned the risk of extinction a probability of greater than 10 per cent within 10 years. This is no longer science fiction.
Also Read: Under Bhagwat too, Hindutva has a two-pronged objective—Hindu supremacy, Muslim subordination
The good, the bad, the ugly
But the news is not all bad. AI is basically a double-edged sword. It can increase human productivity, but it can also causes a lot of damage in its wake.
In a seminar I co-direct at Brown University, Archon Fung, a political theorist at Harvard Kennedy School, brought out the dualistic nature of AI well. He probed AI’s impact in three realms: workplace, social life and politics. And he argued that AI can empower human beings; it can replace them; and it can also subordinate them.
For reasons of space, let me concentrate on the workplace, and especially on the examples of replacement and empowerment.
Fung argued that he does not need to recruit students as research assistants any more, for AI can easily perform that task. In my own research life, this is beginning to happen. I now need fewer assistants for conducting background research for my own writings. AI is a supremely efficient research assistant. (In journalism, managing editors might also soon find that sub-editors are not needed any more. AI can “sub” quickly and well.)
In addition, AI can also write term papers, do problem sets and solve mathematical equations. Papers and take-home exams are being increasingly replaced by the old-fashioned in-class handwritten exams. Musicians are beginning to say that if instructed well, AI can sing and play.
These are all examples of replacement – in full or in part. But what about empowerment?
Writings codes has been called one of the most empowering functions of AI. Of course, not everyone can, or will, use codes. Are there examples of AI making everyday life easier?
Fung writes: “Millions of people have been empowered by AI in their various quests to learn and search, respond to their most important incoming emails and organise their schedules, perfect pursuits like cooking or home maintenance, book travel for business or vacations, and countless other tasks.”
This is clearly life made easy, or made many times easier.
What are the implications of all this for India? AI’s potential to simplify life seems largely inescapable, and a lot of it will soon begin to happen in middle-class homes. A tricky issue will be how to handle replacement. The work of delivery boys can easily disappear in bigger cities. Labour replacing functions of AI can also hit agriculture in rural India.
But none of this is inevitable.
Assessing the relative worth of potential benefits and losses, government policy can play a big role in picking the appropriate level of AI use. The same principle would apply to defence and security – in determining the extent to which AI will be used for weaponry and battlefield calculations. In short, political choices will remain critical to how India pursues AI.
Ashutosh Varshney is Sol Goldman Professor of International Studies and the Social Sciences and Professor of Political Science at Brown University. Views are personal.
(Edited by Insha Jalil Waziri)