I compared the privacy of ChatGPT, Gemini, Claude and Perplexity — here’s the one you should trust most with your personal info
The article reveals significant disparities in how major AI chatbots handle user data, highlighting that privacy is not a standard feature across the industry. It concludes that Anthropic’s Claude offers the highest level of privacy by default, as it does not use conversations for model training and lacks a routine human review pipeline. This stands in stark contrast to competitors like OpenAI’s ChatGPT, which defaults to saving history indefinitely and training models on user interactions unless users manually intervene. The review underscores the importance of understanding platform-specific data retention policies, noting that even services with robust control panels, such as Google’s Gemini, may expose data to human reviewers if settings are not carefully configured. While Perplexity provides granular user controls to opt out of training, its default behaviors still require active management to ensure complete confidentiality. These findings demonstrate that users cannot assume privacy; they must actively audit and adjust settings to protect sensitive information from being retained or utilized for commercial purposes. This analysis is highly relevant to the open data community as it exposes the opacity of proprietary AI data practices. It illustrates how the widespread use of user-generated content to train commercial models often conflicts with principles of data sovereignty and individual rights. By revealing these hidden data pipelines, the article advocates for greater transparency and provides a framework for evaluating the ethical implications of data collection, encouraging a shift toward tools that prioritize user consent and minimal data retention.
Source: tomsguide.comPublished on 2025-12-08