‘DeepTok’: The New TikTok Dilemma—DeepSeek is Strong, But Not Secure
- Advisor@AegisIntel.ai
- Jan 29
- 4 min read
DeepSeek: The TikTok of AI Platforms: A Phenomenal Rise Amidst Security Concerns
DeepSeek’s emergence on the global stage as a leading AI platform has drawn comparisons to the meteoric rise of TikTok. It quickly established itself as a major force in the AI space with the rapid release of its flagship DeepSeek-R1 model just last week. Much like the controversial social media app, DeepSeek attracts users with its advanced technology and seamless user experience. However, as with TikTok, serious concerns over security, transparency, and data integrity have begun to overshadow the platform’s incredible growth.
Meteoric Rise
DeepSeek R1 burst onto the scene with a groundbreaking model that quickly outperformed even the foremost US AI firms. Hundreds of billions in USD and several years of epic development were cast aside on Monday, grabbing global attention.
"Disruption" is a kind word. The impact on financial markets this week has been nothing short of seismic, due to DeepSeek’s novel engineering approaches and their implications for current market leaders. By delivering comparable results to leading AI competitors at a fraction of the cost, its cutting-edge technology became widely adopted, topping download charts and gaining a strong user base, hitting #1 in the Apple App Store. The app is already available via Perplexity AI, and has distilled multiple smaller versions, some small enough to run on mobile devices. Mass distribution accomplished.
DeepSeek’s AI capabilities are undeniably strong. By leveraging multimodal approaches—integrating text, image, and video data—it creates a robust, efficient model that competes with some of the most expensive and resource-intensive systems from the U.S. But the platform’s Chinese roots have raised critical questions about its long-term reliability. Issues of censorship, lack of transparency in data handling, and potential government oversight are already troubling analysts, executives, and government leaders alike.
The Security Flaws
Much like TikTok, DeepSeek is coming under scrutiny for its security practices. While TikTok’s ownership has led to concerns over user data access and government influence, DeepSeek faces similar accusations. Critics point out that DeepSeek’s integration of AI with vast data processing capabilities leaves users exposed. The platform’s ability to handle massive amounts of data—unstructured and structured—at unprecedented speed is both its biggest strength and a potential security liability.
Top 10 Concerns on DeepSeek Security
Lack of transparency in how user data is stored and managed
Risk of government influence, including compliance with Chinese data-sharing laws (confirmed, e.g., Tiananmen Square censorship in the model)
Insufficient evidence of robust encryption or other technical safeguards
Limited visibility into the data governance framework and internal policies
Vulnerability to cyberattacks or external breaches: DDoS attacks on their open-source platform, as well as earlier incursions, leave enormous gaps for data poisoning, backdoor attacks, privacy breaches, malicious code generation, and advanced persistent threat insertion
Ambiguity in data retention practices and how long user information is kept—DeepSeek can retain user data even after cancellation
Reports of censorship mechanisms that could affect the integrity of user interactions
Uncertainty about the extent of third-party access to sensitive user data
Concerns over whether updates or changes to the platform’s AI models introduce new security gaps
The platform’s rapid adoption may outpace its ability to implement effective security measures
Data Privacy Issues—Another Trojan Horse in the Making?
DeepSeek's Terms of Service and Privacy Policy outline the platform's data usage practices. According to these documents, DeepSeek collects various types of user data, including:
Text or audio inputs
Prompts
Uploaded files
Feedback
Chat history
Other content provided by users
Device model
Operating system
Keystroke patterns or rhythms
IP address
System language
Right to Use
This information is stored on secure servers located within the People's Republic of China. The platform states that it may use this data to improve its services, develop new features, and ensure compliance with applicable laws and regulations. However, given that DeepSeek is based in China, user data may be subject to Chinese cybersecurity laws, which could require the company to share data with government authorities upon request. DeepSeek's logging of keystrokes implies enormous concerns.
User data is captured and kept by DeepSeek in China. (It remains to be seen whether running the model locally can mitigate this issue. Although logically sound, this is unlikely to benefit most users due to the necessary computational requirements. Given the data breaches mentioned above, a secure codebase is still unproven. (Go Here for a detailed & authoritative assessment)
Moreover, the terms of service do not clearly grant users rights to the model outputs, potentially opening the door for legal claims by DeepSeek on startup ideas generated using the model. Worse, users are limited to seeking redress only through Chinese courts, while the company retains user data even after account deletion.
DeepSeek’s Terms of Service outline the rights and responsibilities of users concerning the inputs they provide, and the outputs generated by the platform.
According to the DeepSeek Open Platform Terms of Service, users retain ownership of their inputs and are granted rights to the outputs produced by the service. Specifically, the terms state:
Inputs: Users retain any rights, title, and interests in the inputs they submit to the platform
Outputs: DeepSeek assigns any rights, title, and interests in the outputs generated by the services to the user
The New Honeypot?
Will DeepSeek now become the next Chinese-based tech firm to widely access and retain user data from around the world? With a model capable of processing vast amounts of input data in real time, many worry it could become another “honeypot” for user information—an appealing target not only for advertisers and developers, but potentially for state actors. As a Chinese-based tech company, DeepSeek is subject to domestic regulations that might require it to store and share data within its home jurisdiction. This raises the spectre of whether, like TikTok and other Chinese platforms, DeepSeek could serve as a backchannel for global data flow, creating unease among privacy advocates and policymakers worldwide.
Citations:
Comments