Artificial intelligence
From the basics of AI and how it learns to generative models, agents, everyday use and safety. Read one concept, then follow the links to the technologies around it.
Safety & rules4/5
The risks to know before you trust AI, and the safeguards and laws meant to contain them.
- AI safetyThe research and safeguards, taken together, that aim to keep AI from harming people and make it act as people want.
- GuardrailsSafeguards that check requests, answers and actions so an AI does not cross lines it must not cross.
- Red teamingA test where people play the attacker on purpose to find an AI's weak points and dangerous behavior before release.
- SandboxA separate test space where a program can do anything without harming the real system outside.
- Security vulnerabilityA security hole in a program that a bad actor can use to get in, steal information or damage a system.
- EncryptionA technology that scrambles content so that only someone with the key can read it.
- Personal data and privacyInformation that can identify who you are, such as your name, contact details, face or health records.
- CopyrightThe right of people who make writing, pictures, music or video to decide how their work may be used.
- DeepfakeFake video, photos or voices made by AI that look and sound real.
- AI watermarkA mark added so that writing, images, video or sound can be recognized as made by AI.
- AI regulationThe laws and rules that companies and organizations must follow when they build and use AI.
- Sovereign AIA push by a country to build up AI on its own, with its own data, facilities and models fitted to its own language.
Other fields
More on AI
These concept articles were written by AI and checked again for accuracy before publishing. The English edition is translated from the Korean articles. They leave out things that change quickly, such as prices and rankings. Tell us if something is wrong. Report an errorLast updated