DeepSeek, a new generative AI model from China, is challenging Western chatbots and has affected stock valuations, notably causing Nvidia to lose $600 billion. While it offers efficient AI development, concerns arise over censorship and data privacy, as DeepSeek collects personal information and complies with Chinese regulations. Users can access DeepSeek by downloading it or through the Perplexity AI platform, which mitigates data risks by operating on Western servers, though concerns about U.S. data protection persist.
The Rise of DeepSeek: A New Contender in Generative AI
The landscape of generative artificial intelligence (GenAI) has experienced significant upheaval recently with the emergence of DeepSeek. This innovative AI model from China is proving to be a formidable competitor against Western chatbots, leading to dramatic shifts in the stock market. Notably, Nvidia faced a staggering valuation drop of $600 billion as a result.
Often dubbed the Chinese counterpart to ChatGPT, DeepSeek showcases the potential for quicker and more cost-effective AI model development. Its efficient production has largely contributed to plummeting values of major NASDAQ stocks since the end of January.
Should You Switch from ChatGPT to DeepSeek?
Before making the leap to DeepSeek, you might want to explore its capabilities first. However, it’s crucial to be aware that DeepSeek has faced criticism for censoring sensitive requests in line with Beijing’s regulations. Additionally, it reportedly collects personal and technical information, such as pseudonyms, birth dates, email addresses, phone numbers, and passwords, which are stored on servers located in China. This raises concerns regarding compliance with GDPR standards, and as a result, DeepSeek has already been banned in Italy.
So, how can you utilize DeepSeek while navigating the pitfalls of Chinese censorship and minimizing your data risks?
Accessing DeepSeek Safely
One approach is to download and install DeepSeek on your computer. This option caters to tech-savvy users who can handle open-source software. By accessing platforms like GitHub, you can download versions such as DeepSeek v3 and DeepSeek r1, which are the company’s most advanced models. However, this method requires a certain level of computer expertise and the capacity to follow technical instructions, making it suitable primarily for a specialized audience with robust hardware.
If you’re looking for a more user-friendly option, consider using DeepSeek through an online service such as Perplexity AI. This AI-powered search engine has integrated DeepSeek as one of its offerings since January 27. Simply log into your Perplexity account, initiate a new discussion thread, and you’ll see an option to select the AI model you wish to use, including DeepSeek r1.
This setup eliminates the need for installation or configuration, providing a seamless experience. Moreover, Perplexity hosts its services on data centers located in the United States and Europe, ensuring that your data remains within the confines of Western servers, away from the stringent regulations of China. This also means that the censorship challenges associated with DeepSeek are mitigated.
While some may express concerns about sending personal data to the United States, it’s essential to weigh this against the risks of sending data to China. According to the National Commission on Informatics and Liberty (Cnil), data protection in the U.S. aligns partially with EU standards, though it may not be as robust as in countries like Japan or the UK. Nonetheless, it is certainly a safer alternative compared to the current status of data protection in China.