OpenAI Confirms Prompt Injection as Structural Issue in Large Language Models

OpenAI published a document confirming that prompt injection is a structural issue in large language models (LLMs) rather than a bug. The report acknowledges the risk as inherent to LLM architecture, with no definitive cure but outlines methods to mitigate its causes and effects. No specific technical details, dates, or CVE IDs were provided in the referenced content. The impact involves potential exploitation of LLM-based systems through malicious input manipulation. The article cites OpenAI as the primary source of this assessment.