The recent shuffle between OpenAI, Anthropic, and the US Department of Defense or the Department of War (DoW), as it is now called, brought to light how innovation in technology can have far-reaching effects. It also highlights the classic conundrum of regulation and policy trying to keep up with technology advancement.
In the span of a week, Anthropic rubbed off the DoW the wrong way when it said that the US government shouldn’t be using its tools for “all lawful purposes” since it was not built into the initial contract.
Anthropic is the first AI frontier lab to sign a deal, reportedly worth $200 million, with the DoW for use of its AI tools. However, what Anthropic did not anticipate was the use of its technology for mass surveillance and control of autonomous weapons without any human oversight.
The refusal to back down on the issue led the US government to deem Anthropic a “supply chain risk for national security” and ban it from use by federal agencies.Close on the heels of this, OpenAI signed a similar contract with DoW, claiming the safeguards in place ensure that its technology is not used for the purposes highlighted by Anthropic. However, CEO Sam Altman admitted only days later that the contract opened up OpenAI’s tech for use for mass surveillance and autonomous weapons, pushing it to redraw its contract.
Interestingly, Anthropic’s technology was used for the raid that captured Venezuelan president Nicolas Maduro and more recently, in the war between US and Iran.
The Washington Post reported that the US military’s Maven Smart System, developed by Palantir and powered by Claude from Anthropic, identified and prioritised hundreds of targets in Iran based off data from satellites, surveillance and other sources. The report, citing sources, said that the combination of Maven with Claude issued precise location coordinates and has created a tool to speed up the campaign, “turning weeks-long planning to real-time operations”.
Given the rapid pace of development of AI, it is not surprising that both the companies working on innovation, and the government, are learning as they go. “Companies operating on the frontiers of technology have been in sync with the governments in the past because the government has been a major financier and buyer, for example, for space supremacy or electronics or oil or biotechnology. For the first time we have a technology, which is perhaps the most powerful tech till date, and it has been adopted by the masses first and then it is being considered by the government. This is a unique situation, and it is unsurprising that there are some tussles between the tech companies and the government,” says Pawan Prabhat, co-founder of ShortHills AI.
However, even as Anthropic is back to negotiating its contract with the DoW, the whole issue opens the age-old question on regulation and policies playing catch-up with technology advancement.
This has been witnessed in India too with the advancing pace of digital businesses—such as the Digital Personal Data Protection (DPDP) Act protecting the citizens’ personal data collected by businesses, and the 2024 AI Deployment Advisory advising ethical and safe use of AI.
“Why would you want regulation to run neck-and-neck with technology development? Do you not want the creation of a new regulation to involve involved debates, have a slow, iterative drafting process, and be based on consensus? I am happy with the pace of misalignment, and unless a technology is established and adopted, it is premature to regulate it and stunt its development,” Sajai Singh, partner at JSA Advocates and Solicitors tells Forbes India.
He adds that over-regulation would result in a body of outdated laws and is difficult to legislate as technology is global and has no boundaries, making it difficult to have strict jurisdiction and limits.
However, framing laws around ethical use of technology is a democratic process, requiring participation from key stakeholders, industry, academia, local businesses, in addition to the government. Singh of JSA further highlights biotechnology and quantum computing as upcoming areas requiring examination of how the development will impact society at large, in addition to AI.
Who will bell the cat is the million-dollar question. “For tech regulation, a global effort is good to proactively evaluate risks, harmonise standards and prevent fragmentation, ensuring safety without the reactive chaos,” says Dr Srinivas Padmanabhuni, CTO of Aiensured.
First Published: Mar 25, 2026, 11:55
Subscribe Now