Privacy Rankings of AI Chatbots: Le Chat Leads, Meta AI Lags Behind

Privacy Rankings of AI Chatbots

A new study conducted by data removal service Incogni has shed light on the privacy practices of popular generative AI chatbots. The report, titled "Gen AI and LLM Data Privacy Ranking 2025," evaluated nine AI services using 11 different criteria to measure their data privacy practices 1

1

2

2

.

Top Performers in Privacy Protection

Mistral AI's Le Chat emerged as the most privacy-friendly AI service, earning the top spot in the rankings. The French AI company, founded in 2023, impressed researchers with its limited data collection and high scores on AI-specific privacy issues 1

1

3

3

. Le Chat's privacy-conscious approach includes sharing user prompts only with service providers, setting it apart from many competitors 4

4

.

OpenAI's ChatGPT secured the second position, praised for its clear presentation of privacy policies and transparent communication about data usage. The platform also offers users clear ways to limit the use of their data 1

1

2

2

. Following closely in third place was xAI's Grok, which performed well despite some transparency concerns 3

3

.

Mid-Range Performers

Anthropic's Claude and Inflection AI's Pi rounded out the top five, demonstrating relatively good respect for user privacy despite some areas of concern 1

1

2

2

. DeepSeek took the sixth spot, falling in the middle of the pack 1

1

.

Privacy Challenges for Big Tech

Interestingly, AI platforms developed by major tech companies ranked lower in privacy protection. Microsoft's Copilot and Google's Gemini placed seventh and eighth, respectively 1

1

2

2

. Copilot received the lowest score based on AI-specific criteria, such as data used for model training and the use of user conversations in training 1

1

.

The Bottom of the List

Meta AI, developed by Facebook's parent company Meta, ranked last and was deemed the least privacy-friendly AI service among those studied. It received the worst grade for overall data collection and sharing practices 1

1

2

2

3

3

.

Key Findings on Data Sharing and User Control

The study revealed that AI companies share data with various parties, including service providers, law enforcement, corporate group members, research partners, and third parties. Some concerning practices were highlighted, such as Microsoft's policy implying that user prompts may be shared with third-party advertisers 1

1

2

2

.

User control over data usage varies significantly among the services. ChatGPT, Copilot, Mistral AI, and Grok allow users to prevent their prompts from being used for model training. However, Gemini, DeepSeek, Pi AI, and Meta AI do not seem to offer this option 1

1

2

2

.

Transparency and Readability of Privacy Policies

The research emphasized the importance of transparent and readable privacy policies. Companies like Microsoft, Meta, and Google were criticized for having a single, lengthy privacy policy covering all their products, making it difficult for users to find specific information 1

1

2

2

.

Implications for Users and the Industry

This study highlights the growing importance of privacy considerations in the rapidly evolving field of generative AI. As these technologies become more integrated into daily life, users are encouraged to be more aware of how their data is being collected, used, and shared 3

3

4

4

. The findings may also push AI companies to improve their privacy practices to remain competitive and trustworthy in the eyes of privacy-conscious users.