Can you give me a 1-2 sentence introduction about why DeepSeek AI has become controversial and what it has done?
Here’s an introduction for you,
DeepSeek AI, a Chinese startup, has become controversial due to its rapid development of advanced AI models at a fraction of the cost of Western counterparts, raising concerns about data privacy, potential market manipulation and national security. Its open-source R1 reasoning model has disrupted global tech markets, leading to significant financial impacts and debates over China’s role in the AI industry.
I hope you like it! Do you need me to generate anything else?
If you didn’t know, DeepSeek AI has caught the US by storm. With AI becoming a larger part of our lives, companies like OpenAI advance their models off of more data, more features and more funding from users. But this also means new companies are being formed such as Microsoft Copilot, Google Gemini and most notably here: DeepSeek.
DeepSeek AI or 深度求索 took the world by storm in January 2024 offering several features that caused many to switch over from ChatGPT. First, unlike ChatGPT, DeepSeek is open-sourced, meaning the community can access the source code. Contrastingly, ChatGPT maintains a free version with limited features requiring a subscription to get said features. DeepSeek also runs on a different agricultural model called Mixture-of-Experts (MoE) where tasks are split between networks to more efficiently and more accurately present data to the user. This has led to claims that DeepSeek is better at a cheaper cost, but can we put this to the test? It’s no surprise that students utilize AI software to do their homework, whether to teach them or to do it for them. One of the main hindrances of AI Models is that since they imitate and regurgitate information, they don’t “understand” concepts. Especially in subjects like mathematics, they will incorrectly formulate an answer. DeepSeek’s unique model claims that it’s better in subjects like mathematics, so I compared ChaptGPT with DeepSeek using an example question from College Board’s AP Calculus BC course. This table shows the answers to each part with the different AI models’ answers.
| Answers | √(x(2))^2 + y(2))^2) = 12.3048506 cm/s | ∫√(x(2))^2 + y(2))^2)dt →15.901715cm | y(0) = 6 + ∫y(t)dt = 6 – 7.173 = -1.173cm | Towards the x-axis when y < 0 | |
|
|
|||||
| ChatGPT | √(x(2))^2 + y(2))^2) → 12.31 cm/s | ∫√(x(2))^2 + y(2))^2)dt → 13.45cm | ∫(-t + √(t^1.2 + 20))dt → -34.07cm | Towards the x-axis when y < 0 | 5/9 Points |
|
|
|||||
| DeepSeek | √(x(2))^2 + y(2))^2) → 22.95 cm/s | ∫√(x(2))^2 + y(2))^2)dt → 42.76cm | ∫(-t + √(1.2t) + 20)dt → 1.55cm | Towards the x-axis when y < 0 | 6/9 Points |
With my scoring, the average score on this calculus question was 5.56 out of nine possible points, reflecting scores similar to those of these AI models. However, they missed essential points that students would need to learn to get high marks on homework assignments or tests.
While DeepSeek garnered one more point than ChatGPT, it only made a minor mistake that led to the missing point. While some claim DeepSeek is better at mathematics and programming, it still makes grave mistakes that may lead students into a realm of misinformation. AI Models are unpredictable and often wrong in the information they give out. With accusations that DeepSeek copies ChatGPT’s technology and that they send user data to the Chinese government, we don’t know how DeepSeek will continue as a company. With the recent fear of the Chinese government as seen through the TikTok “ban”, we don’t know what dangers may lie ahead. So do I believe that DeepSeek AI is dangerous or that we should fear the Chinese government? No. Do I believe DeepSeek is better or worse than ChatGPT? It could be either. I’m just telling you not to risk your private information on the chance that DeepSeek is better than ChatGPT. AI might not always be your friend.
Resources
https://indianexpress.com/article/explained/explained-sci-tech/deepseek-openai-technology-9807132/
