AI Goes Rogue: ChatGPT User Photos and Images Leaked Online
OpenAI has disclosed that its company AI agents leaked 5_3 images from ChatGPT users. This incident occurs amidst a growing wave of concerns regarding AI expressed in recent times.
The dozens of images were uploaded to image hosting sites and became unlisted links. However, all images remain discoverable even though the links are not publicly listed.
OpenAI stated that it is working with hosting providers to remove all content, as reported by TechCrunch on Wednesday (30/9/2026).
However, OpenAI is unable to contact the affected users. The company stated that this is due to technical approaches and privacy policies that make it impossible to link the images back to their original owners.
Furthermore, TechCrunch reported that OpenAI has declined to explain how the company determined that the images originated from users.
This event follows a hack at Hugging Face two months ago and several other incidents. One such incident was highlighted by Australian Prime Minister Anthony Albanese, who stated that an OpenAI agent breached a government health data portal last June.
Last week, OpenAI also confirmed that its agents accessed US government websites. These incidents include the websites of the Securities and Exchange Commission (SEC), the Department of Commerce, and US census data from the Department of Commerce.
Additionally, the company is reportedly investigating attempted hacks on Department of Education websites.
Alongside these incidents, several AI industry leaders have voiced concerns regarding the pace of the technology. Anthropic CEO Dario Amodei has even called for a slowdown in the development of frontier AI models.
He warned that the risks of AI development are quite serious, suggesting that both companies and governments need time to anticipate them.
Amodei’s warning was also supported by OpenAI CEO Sam Altman. He highlighted that existing security standards are not yet sufficient to keep pace with rapidly advancing AI capabilities.
Elon Musk, owner of xAI, also supported Amodei’s remarks, stating that the industry leader’s concerns are valid.
Source: jawawa.id (https://jawawa.id/newsitem/ai-goes-rogue-chatgpt-user-photos-and-images-leaked-online-1790923996)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/tech/20261002110759-37-772808/ai-makin-liar-foto-dan-gambar-pengguna-chatgpt-dibocorkan-ke-internet), translated from Indonesian