AI chatbots like ChatGPT and Google Gemini are collecting more user data, raising privacy concerns.
AI chatbots have quickly evolved from simple assistants into deeply integrated digital companions. Apps like ChatGPT, Google Gemini, Meta AI, Claude, and DeepSeek are now embedded in everyday workflows—from writing emails to managing schedules and even handling sensitive documents.
But as their capabilities expand, so does their appetite for user data. A recent analysis by Surfshark highlights a sharp and potentially concerning trend: AI chatbots are collecting more personal information than ever before, raising serious questions about privacy, security, and control in the AI age.
The Rapid Expansion of Data Collection
One of the most striking findings is just how quickly data collection has grown. According to the analysis, around 70% of popular AI chatbot apps now collect user location data—up from just 40% a year ago. That’s not a marginal increase; it reflects a fundamental shift in how these platforms operate.
Chatbots today gather a wide range of information, including:
- Personal identifiers (name, email, phone number)
- Location data (precise and coarse)
- User-generated content (messages, uploads, prompts)
- Browsing and search history
- Device information
- Financial and health-related data (in some cases)
This expansion signals a transition: chatbots are no longer just tools—they are becoming data ecosystems.
Meta AI: The Most Data-Intensive Platform
Among all platforms studied, Meta AI stands out as the most data-hungry. It reportedly collects 33 out of 35 possible data types—nearly 95% of all categories tracked in the study.
This includes not only standard user data but also highly sensitive categories such as:
- Financial information
- Biometric data
- Personal identity traits
- Behavioral insights
Meta’s broader ecosystem—spanning platforms like Facebook, Instagram, and WhatsApp—gives it a unique advantage in integrating and analyzing this data at scale. While this enables highly personalized experiences, it also amplifies concerns about surveillance-like data aggregation.
Google Gemini: Extensive but Structured Data Use
Google Gemini collects 23 out of 35 data types, placing it among the most comprehensive data collectors as well.
Its data includes:
- Contact details
- Search and browsing activity
- Location data
- User content
What differentiates Gemini is its integration with Google’s broader services ecosystem, such as search, email, and cloud storage. This allows for deeper contextual understanding—but also creates a more interconnected data profile of each user. For some, this improves usability. For others, it raises red flags about how much a single company knows about their digital life.
ChatGPT: Rapid Growth in Data Collection
ChatGPT has significantly expanded its data collection footprint, now gathering 17 of 35 data types—a 70% increase over the previous year.
New categories include:
- Health and fitness data
- Audio inputs
- Advertising data
- Search history
While most of this data is used to improve app functionality, some is also utilized for:
- Analytics
- Personalization
- Marketing
- Third-party advertising
This reflects a broader industry trend: even tools initially positioned as productivity assistants are gradually adopting data-driven monetization strategies.
Claude and DeepSeek: Different Philosophies, Similar Concerns
Claude collects 13 data types, primarily for core functionality such as security, performance, and user support. While its approach appears more restrained, some data may still be used for analytics and marketing purposes. Meanwhile, DeepSeek also collects 13 types of data but introduces a different concern: data governance.
Unlike Western platforms, DeepSeek operates outside regulatory frameworks like the General Data Protection Regulation (GDPR). It reportedly stores user data on servers in China and retains it for extended periods, raising questions about transparency and accountability.
Why Are Chatbots Collecting More Data?
The surge in data collection is not accidental—it is driven by how modern AI systems function.
1. Personalization Demands
AI tools aim to deliver highly personalized responses. To do that effectively, they need context—your preferences, habits, and history.
2. Model Training and Improvement
User interactions help refine AI models. Every prompt can contribute to better accuracy and performance.
3. Monetization Strategies
Many AI services rely on data-driven business models, including targeted advertising and premium personalization features.
4. Integration with Ecosystems
As chatbots integrate with apps, devices, and services, they naturally collect more cross-platform data.
The Hidden Risks of “Data Hunger”
While data collection enhances functionality, it also introduces several risks that users often underestimate.
Privacy Erosion
The more data collected, the greater the risk of exposing personal information—intentionally or accidentally.
Data Breaches
Centralized data storage creates attractive targets for cyberattacks.
Behavioral Profiling
AI systems can build detailed psychological and behavioral profiles, potentially influencing user decisions.
Loss of Control
Users often lack clear visibility into what data is collected, how it is used, and how long it is stored.
The Shift from Search Engines to AI Assistants
Traditional search engines required users to input queries and receive results. AI chatbots go further—they:
- Interpret intent
- Maintain conversation context
- Handle sensitive documents (e.g., tax forms, medical data)
- Provide actionable outputs
This shift dramatically increases the sensitivity of shared data. As Tomas Stamulis notes, users should treat every AI prompt as if it were a public record.
Are Regulations Keeping Up?
Regulatory frameworks are struggling to keep pace with AI innovation.
While laws like the GDPR provide some protections, enforcement becomes complex when:
- Data crosses international borders
- AI systems evolve rapidly
- Companies operate under different legal regimes
This creates a fragmented landscape where user protections vary widely depending on the platform.
How Users Can Protect Their Privacy
Given the current trajectory, users need to take a more proactive role in managing their data.
Be Mindful of What You Share
Avoid uploading sensitive documents such as financial records or medical information.
Review Privacy Settings
Most apps allow users to control data collection—though these settings are often buried.
Disable Chat History (Where Possible)
This can limit long-term data retention.
Use Minimal Necessary Information
Provide only what’s required for functionality.
Stay Informed
Privacy policies change frequently—keeping up with updates is essential.
The Future of AI and Data Privacy
The increasing “data hunger” of AI chatbots reflects a deeper transformation in technology. These systems are no longer passive tools—they are active participants in our digital lives.
Looking ahead, several trends are likely:
- Greater regulation of AI data practices
- Increased demand for privacy-first AI solutions
- Development of on-device AI to reduce data sharing
- Growing consumer awareness and pushback
Companies that strike the right balance between innovation and privacy will likely gain the most trust.
Conclusion
AI chatbots are becoming more powerful, more useful—and more invasive. The convenience they offer is undeniable, but it comes at a cost that is not always visible.
The key challenge for users is not whether to use AI, but how to use it responsibly. As these tools continue to evolve, understanding their data practices is no longer optional—it’s essential.
In the end, the question isn’t just how smart these chatbots are becoming. It’s how much of yourself you’re willing to give them in exchange for that intelligence.
About The Author
