topic Re: DeepSeek R1 Exposed: Security Flaws in China’s AI Model in Threats

DeepSeek R1 Exposed: Security Flaws in China’s AI Model

Kyaw_Myo_Oo — Wed, 29 Jan 2025 04:08:41 GMT

Dear all,

DeepSeek R1, the latest AI model to emerge from China, is making waves in the tech world. Touted as a breakthrough in reasoning capabilities, it has sparked excitement across industries and even impacted AI-linked stocks globally. With its ability to tackle complex problems in math, coding, and logic, DeepSeek R1 is being positioned as a challenger to AI giants like OpenAI.

But behind the hype lies a more troubling story. DeepSeek R1’s remarkable capabilities have made it a focus of global attention, but such innovation comes with significant risks. While it stands as a strong competitor in the generative AI space, its vulnerabilities cannot be ignored.

https://www.kelacyber.com/blog/deepseek-r1-security-flaws/

Questions for discussion: What are the potential societal and ethical implications of DeepSeek's advancements in AI? What are some potential future directions for the company and the AI field as a whole?

I encourage everyone to share your thoughts and insights. Thanks in advance.

Re: DeepSeek R1 Exposed: Security Flaws in China’s AI Model

Caute_cautim — Wed, 29 Jan 2025 21:59:31 GMT

@Kyaw_Myo_Oo

Great question:

There is evidence DeepSeek, distilled the knowledge out of Open-AI's models, breaching its terms of use and infringing on its intellectual property.

According to IBM "Knowledge distillation is a machine learning technique that aims to transfer the learning's of a large pre-trained model, the 'teacher model', to a smaller 'student model'. Its used in deep learning as a form of model compression and knowledge transfer, particularly for massive deep neural networks.

The goal of knowledge distillation is to train a more compact model to mimic a larger, more complex model. Whereas the objective in conventional deep learning is to train an artificial neural network to bring its predictions closer to the output examples provided in a training data set, the primary objective in distilling knowledge is to train the student network to match the predictions made by the teacher network."

Now look at the privacy conditions for DeepSeek, all data resides in China.

Then do your homework and look at the OpenAI's terms of use and now await the fireworks, legal battle and politicians wading in to sort the whole mess out.

If they manage to provide the distillation, it might be a violation of OpenAI's terms of use, and they might take legal action against DeepSeek. If it happens, this will be a long and challenging litigation process, plus remember OpenAI is based in USA and DeepSeek is based in China.

Given now that the Italian Data Protection Authority has requested information from DeepSeek, citing risks to the data of millions of people in Italy. It seems DeepSeek might not last long at all in the EU.

Do your own due diligence and do not readily accept new technology openly without investigating and protecting yourself first.

Regards

Caute_Cautim

Re: DeepSeek R1 Exposed: Security Flaws in China’s AI Model

Kyaw_Myo_Oo — Thu, 30 Jan 2025 13:10:30 GMT

Hello @Caute_cautim ,

It's great to see you back! I've missed your posts and insightful feedback, always appreciate your contributions to the forum. Thanks for sharing.