Designing AI agents to resist prompt injection

Company news OpenAI Source language: English

How ChatGPT defends against prompt injection and social engineering by constraining risky actions and protecting sensitive data in agent workflows.

Excerpt

How ChatGPT defends against prompt injection and social engineering by constraining risky actions and protecting sensitive data in agent workflows.

Short excerpt from the source — open the original for the full text.

Sources

← Back to the daily brief