Does DeepSeek Track Your Keyboard Input? A Serious Privacy Concern
Image Credit: Towfiqu Barbhuiya | Splash
On February 9, 2025, South Korea’s National Intelligence Service (NIS) issued a warning about DeepSeek, a Chinese AI chatbot, citing serious concerns over its data collection practices and inconsistent responses. The agency advised government bodies to exercise caution, highlighting security risks related to excessive data gathering and potential information manipulation.
[Read More: DeepSeek AI Faces Security and Privacy Backlash Amid OpenAI Data Theft Allegations]
Excessive Data Collection: A Major Privacy Risk?
According to the NIS, DeepSeek goes beyond typical data collection methods employed by generative AI systems. Unlike other AI services, the app allegedly tracks users' keyboard input patterns, a technique that could allow it to identify individuals based on their typing behaviour. Furthermore, the intelligence agency claims that DeepSeek transfers these records to Chinese servers, including those operated by volceapplog.com, raising fears of unauthorized access by Chinese authorities.
The NIS specifically pointed out that DeepSeek allows advertisers unlimited access to user data, a practice that could lead to invasive tracking and potential misuse of personal information. Under Chinese cybersecurity laws, data stored in China is legally accessible to the Chinese government upon request, further intensifying privacy concerns. These findings have prompted South Korean ministries to block access to the app in certain departments, following similar moves by Australia and Taiwan.
[Read More: Italy Bans DeepSeek AI: First Nation to Block China’s AI Over Privacy Issues]
Can DeepSeek Capture Passwords?
If it Logs All Keyboard Input (Keylogging):
If DeepSeek operates like a keylogger, recording every keystroke, then yes, it could capture passwords entered by users, including those typed into login fields.
This would be a severe security breach, as it would allow access to sensitive credentials.
If it Tracks Typing Patterns Without Capturing Full Inputs:
Some AI services use typing pattern analysis to create behavioral biometrics, meaning it studies how users type (speed, rhythm, pressure) rather than what they type.
If DeepSeek only analyzes patterns without storing the actual characters, it might not capture full passwords. However, it could still infer identity based on unique typing styles.
If Password Fields Are Exempted from Tracking:
Secure applications often disable tracking in password fields (such as websites using HTTPS and secured input forms).
If DeepSeek follows such security practices, it may not record passwords directly.
Even if DeepSeek doesn’t explicitly record passwords, the fact that it collects keystroke data means it could potentially reconstruct user behaviour and compromise security through AI-powered inference.
What Should Users Do?
Avoid typing passwords or sensitive information while using DeepSeek or any app suspected of tracking keyboard input.
Use password managers that autofill login credentials instead of typing them manually.
Enable two-factor authentication (2FA) to add an extra layer of security.
Monitor account activity for unusual login attempts.
[Read More: DeepSeek AI Banned in Multiple Countries Over Data Privacy & Security Concerns]
About The National Intelligence Service
The National Intelligence Service (NIS) is South Korea's primary intelligence agency, established in 1961 as the Korean Central Intelligence Agency (KCIA). Over the years, it has undergone several transformations, including a renaming to the Agency for National Security Planning (ANSP) in 1981, before adopting its current name in 1999. The NIS is tasked with collecting and coordinating national security intelligence, with a significant focus on North Korea.
Historically, the NIS has faced criticism for its involvement in domestic politics and human rights abuses. Notably, during the 2012 presidential election, the agency was implicated in an online campaign against the opposition candidate, leading to legal actions against former officials.
In recent years, the NIS has been working to restore its credibility and effectiveness. Under the leadership of Director Cho Tae-yong, appointed in 2024, the agency has sought to counter regional threats and rebuild public trust. However, challenges persist, including internal factionalism and political misuse, which have undermined its operations and public confidence.
[Read More: Navigating Privacy: The Battle Over AI Training and User Data in the EU]
Source: The Korea Herald, Wikipedia, The Guardian, Human Rights Watch, Financial Times
