AI Chatbots have quickly become a staple of life for people around the world. As AI Language Models have become more and more efficient, they have begun to be more integrated into the web, and market. Following the release of ChatGPT in late 2022, the global interest and accuracy of language models spiked, leading to many different competitors releasing models of their own.
In late January, a new AI chatbot model, DeepSeek R1, was released to the public. While a DeepSeek LLM (Large Language Model) has been in development since 2023, the DeepSeek R1 instantly gained popularity because of its response accuracy, especially for the price of a computer it was trained on. Combined with the explicit bias of the AI, DeepSeek made a lasting impression on both consumers and investors.
One of the most significant effects of the release of DeepSeek AI, is the effect on the AI market caused by the development of DeepSeek. DeepSeek is a realistic alternative to ChatGPT, all while being trained on much inferior GPUs (Graphic Processing Units) and significantly less funding. Compared to the billions of dollars spent on getting better GPUs to train AI models further, DeepSeek utilized different software optimization techniques to create an AI that can rival big names in the US markets.
The release of DeepSeek also caused a significant decrease in investor faith in American AI companies, notably Nvidia, whose stock dropped over 17 percent after the release of DeepSeek R1. While Nvidia stock recovered shortly after, the damage was done, and investors realized that investment in the development of AI software could be equally as important as creating stronger computers. This was particularly relevant, considering DeepSeek was created using 2,000 H800 Nvidia chips, compared to the 10,000 chips used by market competitors.
DeepSeek also had lasting effects on consumers, as well as investors. The DeepSeek app would reach the number one app on the iOS store just after the release of the R1 model. While the DeepSeek AI is largely functional for everyday uses, when prompted about the items controversial for the Chinese Government, like the Tiananmen Square Massacre, or the recognition of Taiwan as a country, DeepSeek would either respond with the position of the Chinese Government or refuse to answer the question. For Americans unfamiliar with some of the positions of the Chinese government, being exposed to them for the first time was a little jarring.
DeepSeek has shifted many aspects of the creation around the development of AI. Having an open source (free to view and download) LLM, only makes the creation of new LLMs easier, and with more market competition, is likely to drive AI developers to focus on improving response accuracy. Other aspects of DeepSeek remain industry standard, like the collection of chats that might contain personal information for training purposes, a practice used by most LLM. Additionally, DeepSeek can be easily manipulated and can respond positively to intentional misinformation. According to a Newsweek audit of the response accuracy, DeepSeeks inability to properly recognize misinformation was poor, tying for 10th among its 10 biggest competitors in terms of ability to recognize misinformation.
Consumers and Investors were not the only parties interested in DeepSeek. In early February, just under a month after the release of the R1 DeepSeek AI, governments began to ban and restrict the use of the DeepSeek app on government devices. Other apps, such as TikTok, are already banned on government devices, suggesting that the government believes that DeepSeek in some way collects user data. Multiple countries and states, including Australia and New York, have introduced legislation or outright banned the DeepSeek app on government devices, citing potential security risk, likely relating to how DeepSeek collects user data.
While it remains unclear whether DeepSeek will see long-term success as a LLM, with the staying power to rival tech giants like OpenAI or Google, it is certain that the release of DeepSeek has had long-term implications for the future of AI development. DeepSeek demonstrated that LLM software is equally as important for the development of increasingly larger AIs capable of more and more things.
