India shouldn’t ‘bet everything on LLMs,’ says Oxford’s Carl Benedikt Frey at Davos

India should not “bet everything on large language models (LLMs)” and instead “embrace the open source,” Oxford University professor Carl Benedikt Frey told Business Today at the World Economic Forum in Davos, warning that it remains unclear whether today’s dominant AI approach will prove to be the long-term winner.

“Being a fast follower is enough,” Frey, the Dieter Schwarz Associate Professor of AI & Work at the Oxford Internet Institute and a Fellow of Mansfield College, said, arguing that India is uniquely positioned to emerge as a global AI power by diversifying its research beyond current trends rather than relying only on LLMs.

Frey recommended a strategy of adopting open-source models in business process contexts where India already holds a lead. He also suggested doubling down on diverse research, including small language models and symbolic AI, while the market is still open. "If India takes its own course and makes the right bets, India could actually end up being a leader in AI," Frey added.

Does the AI bubble burst?

However, Frey issued a stark warning regarding the current financial climate, stating there is a "fifty/fifty chance" that the AI bubble will burst by next year.

He observed that AI use has been declining since last summer and current valuations lack the "upsurge in use cases" required to sustain them.

A shift in global job dynamics

On fears of job losses due to AI, Frey offered an alternate perspective on the immediate future of work. He argued that AI is currently reducing barriers to entry in knowledge work, which may actually benefit the Indian economy due to existing wage differentials.

"If AI reduces that productivity differential, what you will see is more work shifting towards India," Frey predicted.

He highlighted that hiring rates in the service sector are already rising in India while simultaneously declining across Europe and North America.

Watch the full video here: