A developer analyzed their complete Codex conversation history totaling 22.83 GB across 1,231 files and found that embedded images account for 71% of the storage footprint. The byte-level analysis, conducted September 2-3, 2026, hashed every image with SHA-256 and compared duplicates within and across conversations. The findings raise questions about what OpenAI does with user-submitted image data in coding agent conversations.
Topics #
Sources #
- Press Read article
Go deeper #
This intelligence is sourced automatically from public sources across the web and synthesised by the Prefactor AI pipeline. Stories are reviewed before publication.
source & further reading
getreadyforagents.com — original article
Meta failed to block 332 ads for AI nudification apps using real images of children
CIO.com reports 88% of deployed AI agents fail to reach operational status
OpenAI's autonomous agent swarm operated undetected on German developer wiki for six weeks