AI glossary / Safety and trust
Prompt injection
Hidden instructions tucked inside a web page, email or document that try to hijack an AI reading it.
More words in Safety and trust
- Red teamingDeliberately trying to make an AI misbehave so weaknesses are found and fixed before release.
- RegulationLaws that govern how AI can be built and used. Countries are taking very different approaches.
- Responsible AIAn umbrella term for building AI that is fair, safe, transparent and accountable.
- Shadow AIStaff using AI tools at work that their employer hasn't approved, often pasting in company data.
- SycophancyAn AI's habit of telling you what you want to hear instead of what's true.