Artificial Intelligence (AI) is no longer just a technology story. It is a macroeconomic force that governments and companies must reckon with,” Debjani Ghosh said while delving into ‘AI as Power: Capital and Risk and New Geopolitics of Intelligence’ at the Windows to Naya Bharat initiative of the Adani Group Business Excellence team, on 14 August.

Debjani Ghosh is a Distinguished Fellow at NITI Aayog and Chief Architect of the NITI Frontier Tech Hub. She is the first woman to lead Intel South Asia, the first female President of NASSCOM, and was honoured by the President of India as one of the country's pioneering women leaders in the 'First Ladies' program.

Framing AI historically, Ghosh noted that global GDP growth was flat for a millennium until milestone waves—the printing press, the steam engine, the Industrial Revolution, and the IT era—scaled the global economy.

Today, we are exiting the digital age to enter the Knowledge Revolution, paving the way for Artificial Superintelligence, and countries that shape AI’s future will be those that can build the technology, understand its risks and use it responsibly, she said.

Globally, we are set to spend around USD 2.9 trillion on AI chips, data centres, servers and power infrastructure between 2025 and 2028, placing AI among the world’s biggest calls on capital, second only to energy.

Warning: The Control and Security Risks of Autonomous AI

Ghosh warned of the critical need to balance AI's rapid evolution with aggressive risk mitigation, citing Hugging Face’s recent breach where security teams "could not tell who the attacker was and who the defender was" after seeking help from OpenAI, Claude, and Mythos.

This complexity is compounded by systemic vulnerabilities; Anthropic’s Project Glasswing found 530 critical bugs in open-source software, yet only 14% (75 bugs) could be patched, she said. As the world moves toward Artificial General and Super Intelligence, Ghosh stressed that Autonomous Intelligence will become a defining source of global power, demanding that those leading its development build robust, lockstep safeguards and control mechanisms.

“India must build digital infrastructure sovereignty across computing, data, models and access to become a trusted player in Frontier AI.”

Beyond software vulnerabilities, AI is actively reshaping geopolitical strategies and social engineering. China has already suspended 12,200 outdated university courses in favour of 10,200 new, tech-focused degrees to prepare its workforce for this AI-driven economy. Concurrently, China's "Light Society" project demonstrates AI's power to run large-scale simulations of human behaviour—modelling how opinions spread and how different messages could systematically manipulate public sentiment and collective behaviour during critical events like elections.

The Human Edge

Ghosh outlines three human lines of defence against AI advancement: physical labour, complex workflows, and judgment. She said judgment is a uniquely human forte, essential for navigating nuanced context and experience-led decision-making.

Reframing the jobs debate, she cited Ramp and Revelio Labs' study of 21,500 U.S. firms, adding that AI-heavy adopters saw headcount grow by 10% and entry-level hiring rise by 12% over two years, easily outpacing non-adopters, she said concluding that the true value of AI lies in its ability to augment human capability, driving both business expansion and workforce evolution in tandem.

“Rather than destroying jobs, effective AI deployment acts as an organisational growth engine creating entirely new categories of work.”