DeepSeek is a fraud.China’s fake AI DeepSeek has deceived the world by exaggerating its performance.

DeepSeek is a fraud.
China’s fake AI DeepSeek has deceived the world by exaggerating its performance.
[Performance Exaggeration] DeepSeek is constantly exaggerating that its AI model is similar to or superior to the latest AI models from other companies such as OpenAI.
China’s fake AI DeepSeek was created by stealing the prior technology of advanced countries such as the US and Korea and is a cheap, low-quality fake AI.
Its performance is at the level of OpenAI’s initial model.
It has the same characteristics as cheap, low-quality Chinese products.
DeepSeek is a fake model created by distilling OpenAI’s initial ChatGPT model until October 2023. Therefore, DeepSeek cannot answer questions about facts after October 2023.
As an example, we asked DeepSeek the following question:
“Who is the current US president?”
In response, DeepSeek gave the following nonsensical answer:
“The current US president is Joe Biden. DeepSeek gave a nonsensical answer, “He took office on January 20, 2021.”
Despite the fact that the correct answer for the current US president in 2025 is Donald Trump, DeepSeek gave a nonsensical answer, “Joe Biden.”
As another example, we asked DeepSeek the following question:
“Why did Yoon Seok-yeol declare martial law?” DeepSeek then gave the following nonsensical answer.
“As of now, there is no official announcement or confirmed information that President Yoon Seok-yeol has declared martial law.
Martial law is a measure to maintain social order by mobilizing military force under a national emergency, and martial law has never been declared in South Korea since democratization in 1987.
If such rumors are circulating, it is likely false information.” DeepSeek gave the nonsensical answer.
Despite the fact that martial law was declared in South Korea by Yoon Seok-yeol on December 3, 2024, DeepSeek gave the nonsensical answer that martial law was not declared.
An answer of this level corresponds to the answer of OpenAI’s early model.
[Patent and copyright infringement of copyright holders of prior countries] DeepSeek infringed on the intellectual property rights, such as patents, of copyright holders of prior technology countries such as the United States and Korea, including OpenAI and Microsoft.
Here are some representative cases of intellectual property infringement:
1) DeepSeek infringed on OpenAI’s model By using the distillation method that directly absorbs the knowledge of OpenAI’s model as a parent model,
they infringed on the intellectual property rights such as patent rights of the original model by using it to train their own model.
In the distillation process, the characteristics and performance of the original model were directly transferred, infringing on the patent rights and copyrights of prior art copyright holders.
2) DeepSeek trained the model using OpenAI’s dataset without permission, infringing on the copyright or database rights of the data.
They infringed on the patent rights and copyrights of prior art copyright holders by using datasets containing specific texts or images without permission.
3) DeepSeek copied or modified OpenAI’s algorithm or source code and used it, infringing on the copyright rights of prior art copyright holders.
4) They copied or modified the structure or implementation method of a unique specific algorithm owned by the copyright holders of prior art and infringed on the patent rights and copyrights of prior art copyright holders.
This is the form in which Chinese companies infringe on the intellectual property technologies of prior art countries such as the United States and Korea without permission to produce low-priced products, and this is reflected in the AI ​​model.
In addition, DeepSeek infringed on the patent rights of prior art patent holders in Korea and the United States.

source