Chinese AI Models Challenge Silicon Valley's Dominance (Full Transcript)

Speaker 1: No artificial intelligence being developed in China has ignited some panic across Silicon Valley. The Chinese AI lab releasing models that can outperform America's best, despite being built with less powerful chips. The National News desk Jeff Harris joins us Jeff. These developments are raising concerns on whether America's global lead in AI is shrinking. What are experts telling you about this?

Speaker 2: That AI lab in China is called Deep Seek DD. It's a free open source language model late last year in December, and what's concerning the US is that it only took two months and less than $6 million to build the release of this AI, not only raising concerns about whether America's lead in artificial intelligence is beginning to shrink, but also calling into question how much is being spent. Building AI models. CNBC reports that in some testing, Deep Seek outperformed Metas Lama 3.1 and open AI's GPT 40 in accuracy, but an expert we spoke with says he doesn't think Deep Seek is posing

Speaker 3: much of a threat. I think this is an opportunity for organizations to do better because there's something publicly available. Right now, and they can now use it and build a better model on top of it, or

Speaker 2: using it now earlier this week, Deep Seek released another model that also outperformed outperformed open AI's latest model in many of those third party tests. Deep Seek isn't the only Chinese company developing AI either. Ticktocks parent company ByteDance released an update to its model on Wednesday claiming that it also outperforms open AI's most recent model.