Does Human Approval Stop Actions Triggered by Prompt Injection?
Only when the approver sees the true action and destination, and approval is reserved for consequential steps. Check whether injected content shapes it.
Research topic
automation bias
Research on automation bias: AI agent security, tokenization and checks for sensitive data access before allowing enterprise workflows to run.
What Does the EU AI Act Require for Human Oversight of AI Agents?
For high-risk uses, Article 14 requires people who can oversee and stop the system, and Article 26 assigns overseers and six months of logs.