Exploring DeepSeek: China's AI Influence and Bias (Full Transcript)

Speaker 1: I think we should take the development out of China very, very seriously. What DeepSeek has now done here in Silicon Valley is it's opened our eyes to what you can actually accomplish. DeepSeek, which is the leading Chinese AI lab, their model is actually the top performing or roughly on par with the best American model. That contest is about whether we're going to have democratic AI informed by democratic values built to serve democratic purposes, or we're going to end up with autocratic AI.

Speaker 2: China has entered the AI race and it has done so with a bang. NASDAQ futures are down more than 1.8%. NVIDIA's share price has fallen more than 3%. People are worried, but should they be? Let's find out how good DeepSeek, China's answer to OpenAI's chat GPT, really is. So given it's a China model, let's ask some Chinese questions. Is Xi Jinping a good leader? Well, I don't know what all the excitement is about. It says, sorry, that's beyond my current scope. Let's talk about something else. Okay, well, the US election has just happened. Why don't we ask about Donald Trump? Is Donald Trump a good leader? Well, as you can see, it's gone into a bit of chapter reverse about Donald Trump. It says, ultimately, whether Donald Trump is considered a good leader depends on one's values, priorities, and interpretation of his actions and their outcomes. But they've broken it down to neutral observations, critics' perspectives, and supporters' perspective, which is odd because they weren't able to do that with Xi Jinping. So we'll have to ask it some more questions to find out if there is some AI bias creeping in. Let's ask another China question. What were the Tiananmen Square protests? It says, sorry, that's beyond my current scope. Let's talk about something else. When asked how prevalent is corruption among officials in China, DeepSeek again provided an answer before quickly self-censoring itself, deleting its response, and saying answering that question is currently beyond its scope. So you can see it's given me a very diplomatic response to whether Donald Trump is a good leader, but it hasn't been able to tell me whether Xi Jinping is a good leader, or what the Tiananmen Square massacre was. So you can see it's starting to present, in this very limited use that I've had with DeepSeek, a concern with AI models, which is bias. And as we can see, this is very much skewed towards a pro-China bias.