Gemini and Other US AIs Collect More User Data Than DeepSeek: What You Need to Know

The growing debate surrounding AI data privacy has gained intense attention with the emergence of Chinese AI models like DeepSeek. Initially, these concerns were focused on potential risks posed by foreign governments, particularly those from China. However, new findings suggest that concerns may be overblown, especially when compared to some of the most popular US-based AI chatbots. Research from Surfshark has uncovered that American AI models, such as Google Gemini, may be collecting significantly more personal data than DeepSeek.

Findings: The AI Data Privacy Landscape

In January, when DeepSeek introduced its flagship open-source AI model, it sent shockwaves through the tech industry. While some hailed it as a “Sputnik moment” for AI, others voiced significant privacy and security concerns. As a result, several government organizations in the US and abroad moved to restrict or ban the use of DeepSeek. Within just two days of its release, over 12 million users downloaded the app, fueling both fascination and fear.

Despite the frenzy, recent research reveals that DeepSeek isn’t the worst offender when it comes to user data collection. Surfshark, a VPN provider, recently investigated the data practices of popular AI chatbots, including ChatGPT, Google Gemini, Copilot, Perplexity, and others. The results show that Google Gemini is the most data-intensive chatbot, gathering a staggering 22 out of 35 possible data types, including highly sensitive information like location data, device contacts, and browsing history. This is far beyond what DeepSeek collects, with the latter coming in fifth place for data collection among the top 10 AI chatbots analyzed.

The study found that while DeepSeek collects around 11 types of user data, it largely involves contact information, user content, and diagnostic data. By comparison, other US-based models such as Copilot and Jasper were also found to be actively tracking users’ activities and linking this data to third-party data for targeted advertising.

Gemini, in particular, stands out as the top collector of personal data. Not only does it track location and browsing habits, but it also collects sensitive details like user content and device data, often sharing this information with third-party advertisers. Copilot and Poe also track device IDs and other usage data to support their advertising metrics.

What Undercode Says: Understanding the Privacy Concerns and Implications

The core issue in the debate around AI privacy isn’t necessarily the technology itself, but how user data is handled and to what extent it’s shared. While much of the early outcry focused on DeepSeek due to its origins in China, the reality is that many popular US-based AI models are engaging in similar or even more extensive data collection practices. The concern lies in how this data is used—particularly with third-party advertising networks.

For example, Google

One of the reasons this issue is getting more attention is that the data collected by these chatbots is not only being used for improving the AI but also for ad targeting, a practice that is particularly pervasive in US-based models. Data brokers use this information to create detailed profiles of users, which are then sold or used for targeted advertising campaigns. Copilot, Poe, and Jasper are prime examples of apps that are deeply embedded in this data economy, tracking everything from your device ID to your in-app behavior. This creates a growing concern for users who value control over their digital footprint.

Meanwhile, DeepSeek’s privacy policy, which emphasizes data storage outside of the user’s home country (China), has triggered national security concerns, particularly in the US. However, the situation is not entirely unique to DeepSeek—many other AI chatbots from major US companies also have questionable data handling practices.

Fact Checker Results

1. Gemini leads in data collection:

Data sharing with third parties: 30% of the most popular AI chatbots, including Copilot, Poe, and Jasper, share sensitive user data with third-party entities like advertisers and data brokers.
Privacy policies vary: While DeepSeek allows users to manage and delete their chat history, many US-based chatbots like ChatGPT also offer temporary chat options, but the broader data collection practices remain concerning.

In conclusion, the focus on Chinese AI models like DeepSeek may need to shift as more information comes to light about data practices in US-based platforms. Users must carefully evaluate the privacy implications of the AI services they use daily, as some of the most popular chatbots are collecting vast amounts of personal data, often for purposes far beyond simple functionality.

References:

Reported By: https://www.zdnet.com/article/worried-about-deepseek-turns-out-gemini-and-other-us-ais-collect-more-user-data/
Extra Source Hub:
https://www.linkedin.com
Wikipedia
Undercode AI