GitHub Enhances AI Offerings with DeepSeek R1 While OpenAI Examines Possible API Misuse

Microsoft has added DeepSeek R1 to GitHub Models while OpenAI has investigated unusual API access patterns linked to China.

Microsoft has integrated the DeepSeek R1 AI model into GitHub Models, its experimental AI catalog, while OpenAI continues investigating whether proprietary data was accessed without authorization during DeepSeek’s development.

The model, created by China-based DeepSeek, has raised security and compliance concerns as it gains wider availability across Microsoft’s AI platforms.

OpenAI’s security team began reviewing unusual API traffic patterns in late 2024, with multiple developer accounts linked to China exhibiting spikes in requests.

While OpenAI has not publicly confirmed DeepSeek R1’s involvement, concerns have grown over whether outputs from OpenAI models were used to train competing AI systems.

Despite these concerns, Microsoft has now introduced DeepSeek R1 to both Azure AI Foundry and GitHub Models, expanding its accessibility.

DeepSeek R1’s Capabilities and Deployment

DeepSeek R1 is a 671-billion-parameter large language model designed for reasoning tasks, deep learning, and computer vision. According to DeepSeek, the model was trained using 2,048 Nvidia H800 GPUs, hardware that complies with U.S. export restrictions on advanced AI chips. However, doubts have emerged that DeepSeek might have used a higher number and more powerful H100 GPUs.

DeepSeek R1’s efficiency has positioned it as a competitive alternative to OpenAI and Google DeepMind models, requiring fewer computational resources while achieving high reasoning performance.

The model is now available through GitHub Models, where developers can test and integrate it using GitHub’s playground interface or API access.

GitHub Models allows side-by-side comparisons with other AI models, including those from OpenAI, Meta, and Cohere. Microsoft describes this addition as a way to provide developers with diverse AI options, though the decision has sparked concerns over data security and compliance.

Side-by-side comparison of DeepSeek R1 and OpenAI o1 on GitHub Models (Image: Github)

Microsoft continues to expand its AI ecosystem, hosting models from OpenAI, Meta, Mistral, and Cohere on its platforms.

Regulatory and Security Concerns

DeepSeek R1’s growing adoption has triggered multiple regulatory reviews. In Europe, the Italian Garante privacy authority is investigating whether DeepSeek transfers European user data to China without adequate safeguards, a potential violation of the General Data Protection Regulation (GDPR).

In the U.S., the Navy has formally banned DeepSeek AI models from both official and personal use, citing data security risks associated with Chinese AI technologies.

An analysis by NewsGuard found that DeepSeek R1 filters politically sensitive topics, particularly those related to Tiananmen Square and criticisms of China’s leadership. In 85% of tested cases, the model either refused to answer or provided incomplete responses on these topics.

Microsoft has not publicly commented on whether additional moderation measures will be applied before a full-scale rollout of DeepSeek R1 in Azure and GitHub.

AI Market Impact and Microsoft’s Strategy

The efficiency of DeepSeek R1 has contributed to shifts in the AI hardware market. Nvidia’s stock saw a sudden drop of around 17%, with investors reassessing demand for high-end AI GPUs as models like DeepSeek R1 demonstrate the ability to perform complex reasoning tasks with fewer computational resources.

SourceGitHub
Markus Kasanmascheff
Markus Kasanmascheff
Markus has been covering the tech industry for more than 15 years. He is holding a Master´s degree in International Economics and is the founder and managing editor of Winbuzzer.com.
0 0 votes
Article Rating
Subscribe
Notify of
guest
0 Comments
Newest
Oldest Most Voted
Inline Feedbacks
View all comments
0
We would love to hear your opinion! Please comment below.x
()
x