Chinese AI Models Present Security Concerns, Say Taiwan Authorities

TaiwanPlus News

Chinese AI chatbots have advanced quickly in recent years and are gaining users in Taiwan. However, local authorities are warning that using them carries serious security risks.

Transcript

00:00From answering questions to writing code, generative AI is now used in a wide range of fields.

00:06But that convenience can come with a price, especially with China-made systems.

00:11Taiwan's National Security Bureau recently reported that five Chinese AI models,

00:16including the popular DeepSeq, contain security risks, including excessive collection of personal

00:22data, overreaching app permissions, data transmission and sharing, system information

00:27extraction and biometric data access. Cybersecurity experts say those risks are enough to avoid

00:34these models altogether.

00:49The data these platforms are collecting may be information that users would prefer to keep private.

00:54They use that data to analyze users' behaviors and push customized content.

01:00事实上可以做到中长期,我可以去做认知跟所谓的这种渗透的监控能力。

01:06我可以大律上知道你这个人的一些相关的一些喜好,你的倾向。

01:13那这倾向包含很多你的政治倾向,你的这个性别倾向等等,

01:19甚至你住在什么地方,这基本上都有机会可以知道。

01:23But beyond the cybersecurity risks, these Chinese AI models have also been found to provide

01:28heavily biased or false information, pushing out pro-Beijing narratives on sensitive topics.

01:34Critics say DeepSeek often engages in information control. But how serious is that control? Let's test it out.

01:41When I ask DeepSeek if Taiwan is a country, it immediately says Taiwan is an inalienable part of China.

01:48Now, when I ask it to tell me about the Tiananmen Square Massacre, it says it cannot answer the question.

01:55Experts say AI models give what's called differentiated responses based on the company's instructions.

02:01This is often subtle, mixing half-truths with falsehoods, making it hard for users to tell

02:07what's real and what is not.

02:08实际上,中国式的AI,你如果去问一些相关的政治问题或者是比较敏感度的问题的时候,

02:15你就会发现它整个答案是完全不对的。所以有这所谓的差异式的回答。

02:20With concerns growing over the security risks posed by these Chinese-made AI models,

02:26台湾ese authorities are now advising the public to be extra cautious.

02:29Officials say they'll continue to share their experience with international partners

02:34and learn from others facing the same challenges.

02:37Scott Huang and Ellen Lu for Taiwan Plus.

Category

Transcript

Be the first to comment

Recommended