EmailGPT Vulnerability Let Attackers Access Sensitive Data

A new prompt injection vulnerability has been discovered in the EmailGPT service. This API service and Google Chrome plugin help users write emails in Gmail using OpenAI’s GPT model. The prompt injection vulnerability arises when an attacker manipulates a large language model (LLM) using manipulated inputs, allowing the LLM to execute the attacker’s intentions deliberately.  […]

The post EmailGPT Vulnerability Let Attackers Access Sensitive Data appeared first on Cyber Security News.

access api attacker attackers chrome cyber security data emailgpt emails gmail google google chrome gpt injection inputs language large large language model llm openai plugin prompt prompt injection sensitive sensitive data service using vulnerability