Home›Global Markets›Asia›China’s AI race faces a looming shortage of high quali…

China’s AI race faces a looming shortage of high quality training data

Epoch AI estimates the global supply of high quality, publicly available human generated text could be exhausted within six years.

Reported · Written up by our desk · Asia

SCMP Economy reports that China’s push to build next generation AI models is shifting to a less visible but potentially more existential bottleneck, a shortage of high quality training data.

The outlet cites US-based Epoch AI, saying the global supply of high quality, publicly available human generated text could be fully exhausted within the next six years, and that hardware workarounds may not be able to solve the problem.

It also notes warnings from OpenAI co-founder Andrej Karpathy about a looming “data wall” by the end of the decade, which could cause model capabilities to plateau without fresh, reliable information.

SCMP Economy further says the pressure has led major AI labs to spend heavily on mining offline human knowledge, prompting an ethical debate.

More like this

Sources