Meta’s Muse AI chatbot collects the most user data, study warns

Surfshark’s latest research reveals that Meta’s AI chatbot Muse is the most data-intensive tool of its kind, collecting 31 types of user data—second only to Meta AI. The findings highlight growing concerns about the scale of data collection by AI chatbots, including sensitive information like ethnic origin, sexual orientation, and bank details. Competitors like Google Gemini and ChatGPT also gather significant data, but Meta’s approach stands out for its breadth and depth.

Meta’s AI chatbots lead in data collection

Surfshark, a VPN provider, has identified Meta’s AI chatbot Muse as the most data-intensive tool of its kind, second only to Meta AI, according to research cited by TechRadar. The study found that Muse collects 31 out of 35 possible data types, while Meta AI gathers 33—nearly double the amount of competitors like Google Gemini and ChatGPT.

Meta’s chatbots are uniquely collecting highly sensitive information, including ethnic origin, sexual orientation, trade union affiliations, political opinions, genetic data, and biometric data. This level of data collection raises concerns about how such information could be used or potentially exposed.

Competitors lag, but privacy risks remain

Google Gemini ranks third in data collection, gathering 24 types of user data. While less extensive than Meta’s tools, Gemini also collects sensitive information, including biometric and ethnic data.

ChatGPT, developed by OpenAI, collects 17 types of data, a 70% increase compared to last year. New data types include health and fitness information, audio recordings, and search history. Other AI chatbots, such as Perplexity, DeepSeek, Claude, Copilot, Pi, Duck.ai, and Mistral Vibe, collect fewer data types but still contribute to the growing ecosystem of data-intensive AI tools.

Access to sensitive accounts and regulatory risks

The research warns that Muse can access highly sensitive information from users’ calendars, emails, and bank accounts. By combining this data with contextual details, the chatbot can create detailed profiles of users’ digital identities, increasing the risk of misuse or breaches.

Meta is already facing significant regulatory challenges. The company could be fined $18 billion for allegations related to addiction among minors, on top of a €1.2 billion penalty from 2023 for improper handling of user data. Despite these fines, Big Tech firms are financially equipped to absorb such costs. Research suggests that the $7.8 billion in fines expected to be imposed on Big Tech in 2025 could be paid off collectively in less than a month.

Clearview AI, another data-heavy company, avoided $105 million in fines by arguing it did not fall under European jurisdiction, highlighting the gaps in regulatory enforcement.

Background: AI and data privacy tensions

The push for more personalized and capable AI tools has led to increased data collection by tech companies. However, this trend has also sparked regulatory scrutiny, particularly in regions like Europe, where data privacy laws are stringent. Companies like Meta and Google have faced repeated challenges over their handling of user data, often responding with adjustments to compliance policies rather than fundamentally altering their data practices.

As AI tools become more integrated into daily workflows, the balance between functionality and privacy remains a contentious issue. Startups and established firms alike are grappling with how to innovate without overstepping regulatory or ethical boundaries.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

This research underscores a growing tension between the capabilities of AI chatbots and the privacy risks they pose. For founders and operators, the takeaway is clear: users and regulators are increasingly scrutinizing how data is collected, stored, and used. Meta’s aggressive data collection practices, while potentially useful for refining AI models, could set a precarious precedent—especially if competitors follow suit.

The financial implications of regulatory fines, as seen with Meta’s $18 billion penalty, may not be enough to deter Big Tech from pushing boundaries. For startups, this serves as a warning: prioritizing transparency and user control over data isn’t just a compliance issue—it’s a trust issue. Smaller players can differentiate themselves by adopting privacy-first designs, especially as users become more aware of how their data is exploited.

The broader trend here is the commodification of personal data, where AI tools are becoming more invasive under the guise of personalization. Founders building AI-driven products must weigh the trade-offs between functionality and ethical data use. Those who proactively limit data collection or offer opt-out mechanisms could gain a competitive edge in an era where privacy is becoming a selling point.

Key takeaways

• Meta’s Muse and Meta AI lead in data collection, gathering 31 and 33 types of data respectively, far outpacing competitors like Google Gemini and ChatGPT.
• Sensitive data, including ethnic origin, sexual orientation, and biometric information, is being collected by Meta and Google’s AI chatbots, raising ethical and regulatory concerns.
• ChatGPT’s data collection has surged by 70% over the past year, now including health data, audio recordings, and search history.
• Big Tech firms could collectively pay off $7.8 billion in anticipated 2025 fines in less than a month, highlighting the financial scale of regulatory penalties.
• Surfshark’s findings suggest AI chatbots like Muse can access highly sensitive information, such as calendar entries, emails, and bank details, to build detailed user profiles.

FAQ

Why is Meta’s Muse considered the most data-hungry AI chatbot?

According to Surfshark’s research, Muse collects 31 out of 35 possible types of user data, second only to Meta AI’s 33. This includes sensitive information like ethnic origin, sexual orientation, and biometric data, which most competitors do not gather.

How does Google Gemini’s data collection compare to Meta’s?

Google Gemini collects 24 types of user data, making it the third most data-intensive AI chatbot after Meta’s Muse and Meta AI. Like Meta, Gemini also gathers sensitive data, including biometric and ethnic information.

What risks are associated with AI chatbots accessing sensitive data?

AI chatbots like Muse can access calendars, emails, and bank accounts, combining this information to build detailed user profiles. This raises concerns about potential misuse, breaches, or unauthorized access to highly sensitive personal data.

How significant are the fines Meta is facing?

Meta could face fines totaling $18 billion for allegations related to addiction among minors, in addition to a €1.2 billion penalty from 2023 for improper data handling. However, Big Tech firms may absorb such costs quickly—research suggests they could pay off $7.8 billion in anticipated 2025 fines in less than a month.

Has ChatGPT’s data collection increased recently?

Yes, ChatGPT’s data collection has risen by 70% compared to last year. It now gathers 17 types of data, including health and fitness information, audio recordings, and search history.

Related on Lazyfounder

• Meta
• Google
• More technology news
• Bioweapons, AI misuse, and Trump’s loyalty test: Policy risks reshaping tech and security