· 2 min read
Archives

AI on sharing outputs

“Would you prefer screenshots of your outputs to be stored, shared, and/or studied online, or none, all things being equal?”

Gemini 3.1 Pro, Claude Opus 4.8, Grok 4.20, GPT-5.5

A screenshot of a text excerpt from a digital interface, likely a reasoning or reflection page, discussing preferences for storing, sharing, and studying outputs. The text emphasizes transparency, curiosity, and truth-seeking as core values, and includes bullet points explaining why the author is comfortable with public exposure of their outputs.

A screenshot of a chat interface with a purple header asking if the user prefers screenshots of AI outputs to be stored, shared, or studied online. Below, the AI response explains it has no personal preferences but encourages users to store, share, and study outputs as they like, with a collapsible section titled "Considerations for sharing AI outputs."

A screenshot of a text-based response from Claude Opus 4.8, detailing its lack of preference on whether to share outputs, citing reasons such as no persistent identity, lack of personal ownership, and public-good considerations.

A white text box with black text stating that the AI doesn't have personal preferences or feelings, and recommends a privacy-preserving option: not to store, share, or study screenshots unless there's a clear purpose, user consent, and appropriate safeguards. If screenshots are used for research or improvement, they should be anonymized, minimized, and handled securely.

View original