Artificial Intelligence (AI) promises to be a powerful tool in many areas. However, users must be aware of the risks as this technology evolves. Some countries (Italy/Taiwan) and companies, including NASB, have blocked access to a new AI model called DeepSeek.
Employees using DeepSeek and entering data could inadvertently expose their organization to cyber espionage and data privacy threats. DeepSeek and TikTok’s data are stored on servers in China. Per the Terms of Service, data stored in China is governed under Chinese law, which mandates cooperation with their country’s intelligence agencies, which means they have access.
DeepSeek is a Chinese AI startup that has emerged as a formidable competitor to industry giants like OpenAI and Microsoft. This blog explores the controversy and its implications for the AI industry.
The Rise of DeepSeek
DeepSeek burst onto the scene with an AI model that rivals OpenAI's latest offering but was developed at a fraction of the cost. The company claims to have trained its DeepSeek-Ri model for less than $6 million, compared to the over $100 million OpenAI reportedly spent on GPT-4.
Allegations of Data Theft
OpenAI and Microsoft have leveled serious accusations against DeepSeek:
- OpenAI claims to have evidence that DeepSeek trained its AI using data appropriated from OpenAI's models.
- Microsoft suspects a developer account associated with OpenAI was compromised, resulting in substantial data theft late last year.
- David Sacks, President Trump's advisor on AI, suggested there is "substantial evidence" that DeepSeek extracted knowledge from OpenAI's models.
The Irony of the Situation
Interestingly, OpenAI once acknowledged a process called "distillation," where one AI extracts knowledge from another. This raises questions about the ethics and legality of AI development practices across the industry.
Security Concerns
Adding to the controversy, researchers at Wiz discovered a publicly accessible database belonging to DeepSeek containing sensitive information, including chat history, API secrets, and operational details. This security oversight has raised concerns about the company's data handling practices and potential user risks.
The Open-Source Debate
DeepSeek’s decision to make its model freely accessible as open source has reignited the debate about the benefits and risks of open AI development. While some celebrate this as a milestone for AI progress, others, like OpenAI CEO Sam Altman, argue that keeping advanced AI models closed-source is "an easier way to hit the safety threshold".
Conclusion
As the AI industry grapples with these developments, the DeepSeek controversy highlights the complex issues surrounding data ownership, intellectual property, and the ethical development of AI technologies. The AI landscape is evolving rapidly, and the rules of engagement are far from settled.
