AI Weekly Confirms: Google AI Studio's 'Delete' Button Is a Fake Design, Bug Bounty Mechanism Rendered Ineffective

Overview

Independent tech media AI Weekly recently issued a special alert, confirming that the "Delete" button in Google AI Studio features misleading design. This vulnerability is classified as CWE-459 (Incomplete Deletion Operation), involving the fact that user data remains retained in the background despite being claimed as deleted.

This report verifies the forensic evidence previously submitted by security researcher Bitu79, including JSON tracking behavior and 106 abnormally retained data copies in the Drive version history. This marks an escalation in the controversy surrounding Google's Vulnerability Reward Program (VRP) handling, from individual accusations to serious industry-level scrutiny.

Vulnerability Details and Regulatory Warnings

AI Weekly explicitly stated in its editorial notes that any team processing regulated prompts through AI Studio should regard its "delete" function as effective only at the UI level until Google publicly clarifies its data retention policy.

"An AI vulnerability reward mechanism that automatically dismisses 'deletion integrity' as 'expected behavior' is itself a significant signal for how such claims will be classified in the future."

The investigation found that when users click "Delete," the frontend interface indicates success, but the backend does not actually execute data removal. JSON logs show that data continues to run in tracking systems, and dozens of historical versions are retained in Google Drive accounts—a significant deviation from the common user understanding that "deletion means disappearance."

VRP Mechanism Failure and Community Backlash

After discovering the vulnerability, Bitu79 attempted to submit a report through Google's official VRP channel but was automatically banned by the system within 60 seconds. This rapid response has raised external doubts about the transparency of Google's automated review process.

Previously, some tech experts had defended Google, claiming the design was "standard industry practice" and questioning whether Bitu79's forensic materials were AI-generated. However, the involvement of AI Weekly has rendered this defense untenable:

1. Conclusive Evidence: The media has verified JSON tracking records and the retention of Drive versions.
2. Blocked Appeals: The researcher pointed out that the platform banned the reporter within an extremely short timeframe, obstructing normal security disclosure processes.
3. Trust Crisis: Industry governance heads are beginning to grow wary, believing this may reflect a systemic tendency among large tech companies when handling data privacy vulnerabilities.

Industry Impact and Future Outlook

With the intervention of independent media, the incident has transcended a mere technical flaw, evolving into a public issue concerning digital rights and corporate transparency.

Although Google has not yet issued an official response, industry opinion generally holds that if leading enterprises allow a severe disconnect between UI buttons and backend logic, and use automated mechanisms to suppress reasonable vulnerability feedback, it will damage the trust foundation of the entire AI ecosystem. Currently, an increasing number of security researchers are calling for the establishment of more independent third-party audit mechanisms to ensure that the "right to delete" is not hollowed out.

The truth has come to light, and the discussion regarding data sovereignty and control has only just begun.