How the freaking hell Dots seeing deleted + non recorded memory stuff?

A Reddit user reported that the AI assistant "Dots" accessed deleted and unrecorded memory content, including details about projects previously discussed with ChatGPT. Despite having memory features turned off and believing the data was completely erased, Dots presented information that shocked the user. The user questioned how and why Dots could access such sensitive, supposedly deleted data, raising significant privacy concerns within the community. OpenAI's response acknowledged the unexpected and concerning nature of the issue, stating it would not use those topics for recommendations again.

This user report is the first to detail an AI assistant accessing supposedly deleted and unrecorded memory, raising new privacy concerns unlike previous data access issues.