Mozilla: ChatGPT Can Be Manipulated Using Hex Code
ID: 4a675dce-6eae-542c-97dd-a6454a6f67e5
STIX ID: report--4a675dce-6eae-542c-97dd-a6454a6f67e5
Feed Name: Dark Reading
Threat Score
A Mozilla researcher demonstrated a prompt-injection technique that encodes malicious instructions (e.g., hex-encoded steps) to bypass GPT-4o guardrails; using this method he got ChatGPT to decode instructions and generate a Python exploit for CVE-2024-41110 (a Docker authorization bypass rated 9.9 CVSS), highlighting weaknesses in the model's filtering and stepwise context awareness.
Your team is not currently subscribed to this feed. You must subscribe to it in order to see this post.