Billions Committed to AI Safety is Real. The Proof Is Not.

The Problem Is Time, and the Absence of Standards

Most safety initiatives target problems that take years to evaluate. Workforce displacement, misinformation, inequality and societal adaptation cannot be measured in a single funding cycle. The scale of what is being addressed is not in dispute: Goldman Sachs has estimated that generative AI could expose the equivalent of 300 million full-time jobs globally to automation, and the IMF reckons nearly 40% of jobs worldwide could be affected, rising to about 60% in advanced economies. Those forecasts moved the AI debate out of engineering and into economics and politics.

AI companies now describe their own missions in those terms. OpenAI’s latest initiative explicitly raises economic security, wealth distribution, public-ownership models and sovereign wealth funds. “The transition toward superintelligence will come with serious risks, from economic disruption, to misuse in areas like cybersecurity and biology, to the loss of alignment or control over increasingly powerful systems,” the company wrote in its policy proposals. “Without effective mitigation, people will be harmed.”

To Maaz Ansari, CRO and co-founder of customer experience automation firm Oriserve, the framing matters but is not enough. “There has been considerable interest in developing frameworks for implementing AI safety as AI is transitioning from an experimental basis to industry-wide implementation,” he said. The danger is an announcement without follow-through: “The objectives will not be met if they are simply a promise with no measurable impact on society.”

Ankush Sabharwal, founder and CEO of CoRover.ai, a conversational AI firm behind the BharatGPT initiative, put it more sharply. “The rise of AI safety and philanthropic funds from major tech companies is an encouraging signal, but signals should not be mistaken for measurable impact,” he said. Real work is underway in alignment research, red-teaming, model evaluation and responsible deployment, he noted; the trouble is that the safety funding remains far smaller than the capital driving capability.

Why the Accountability Gap Matters More in India

Across these interviews, one theme recurred: there are no widely accepted standards for judging whether AI safety funds work. “The public should move past headline funding amounts and consider measurable outcomes,” Agrawal said. Companies, he argued, should disclose where funding goes, who receives it, what milestones are tracked, and whether independent experts evaluate results: deployment of safety tools, reductions in bias, workforce-training outcomes, energy-efficiency gains. Without those, meaningful progress is hard to tell apart from reputation management.

Ansari said transparency must go beyond financial disclosure to the organizations involved, the outcomes expected, and independent audits. Sabharwal went furthest: trust cannot be built on corporate assurance alone. Accountability, he said, requires public disclosure of governance structures, grant-selection criteria, allocations and independently verified impact, along with a willingness to fund scrutiny even when findings are inconvenient. “If safety research is supported only when it is convenient or reputation-enhancing, people will naturally question its credibility.”

This is where India’s position diverges from the West’s. India is not primarily a funder of AI safety; it is a recipient and a deployment ground. The Gates Foundation and Anthropic’s program names India directly. India’s own institutions, among them the IndiaAI Mission and the recently launched Indian AI Research Organisation, are young. 

The country’s data-protection regime, with rules under the Digital Personal Data Protection Act, or DPDP Act, notified in November 2025, is still moving toward enforcement.

That combination matters. When foreign-funded AI tools are deployed in Indian hospitals, classrooms and welfare systems, India has fewer independent bodies able to verify the safety claims attached to them than the US or the EU does. The accountability gap that worries these executives is, for India, also a sovereignty question.