Jeff Bezos-Backed Perplexity Rolls Out Self-Hosted DeepSeek R1 Model Amid Chinese AI Controversy

In a move that has sparked both excitement and controversy, Jeff Bezos-backed Perplexity AI has rolled out a self-hosted version of the DeepSeek R1 model. The announcement, made by Perplexity CEO Aravind Srinivas on X (formerly Twitter), has drawn attention to the capabilities and potential implications of this advanced AI model.

The DeepSeek R1 model, developed by the Chinese AI startup DeepSeek, is renowned for its powerful reasoning abilities. By integrating this model into Perplexity AI's platform, users can now enjoy enhanced deep web research capabilities and robust data security, as all data is processed exclusively on servers located in the US and EU.

However, the announcement has also raised concerns about the potential involvement of Chinese AI in Western data centers. Ross Gerber, CEO of investment firm Gerber Kawasaki, has criticized DeepSeek's privacy policy, expressing distrust in Chinese data practices. Furthermore, the White House has placed DeepSeek under a national security review, citing concerns about the model's origins and potential risks.

In response to these concerns, Srinivas addressed the misconceptions surrounding the integration of DeepSeek R1 into Perplexity AI. He stated, "All DeepSeek usage in Perplexity is through models hosted in data centers in the USA and Europe. DeepSeek is *open-source*. None of your data goes to China." To reassure users, Srinivas shared a screenshot of an unfiltered response from the DeepSeek R1 model on Perplexity, demonstrating its uncensored nature.

The integration of DeepSeek R1 into Perplexity AI has the potential to significantly impact the competitive landscape of AI search technology, particularly in relation to established models like those from OpenAI. With its cost efficiency, superior performance, open-source nature, and robust data security measures, DeepSeek R1 could challenge the dominance of established models and drive innovation in the AI industry.

As the EU's AI Act approaches implementation in 2026, DeepSeek R1 and similar models may need to adjust their deployment strategies to comply with new regulations. This could involve risk assessments, transparency and explainability measures, robust data governance, ethical considerations, and international cooperation. By proactively addressing these aspects, DeepSeek R1 can ensure its continued use and success in the European market.

In conclusion, the rollout of the self-hosted DeepSeek R1 model by Perplexity AI has sparked both excitement and controversy. While the model's advanced capabilities and data security measures have been praised, concerns about its Chinese origins and potential risks have also been raised. As the AI industry continues to evolve, it is crucial for companies like Perplexity AI to address these concerns and ensure the responsible deployment of cutting-edge AI models.