10 terms
Na when data or AI dey lean one side in a regular way, wey fit make some people or groups get unfair or less correct result.
A systematic skew in data or an AI system that can lead to unfair or less accurate outcomes for some people or groups.
Na when private or restricted information escape by mistake through AI system, training data, or answers.
The unintended exposure of sensitive or restricted information through an AI system, its training data, or its outputs.
Na fake or changed media wey convincingly show person dey talk or do something wey e no talk or do.
Synthetic or altered media that convincingly depicts a person saying or doing something they did not say or do.
Na how well person fit understand why AI produce one particular answer or decision.
How well people can understand why an AI system produced a particular output or decision.
Na rules and technical checks wey dey control wetin AI fit do, to reduce answer or action wey no good or fit cause wahala.
Rules and technical controls that limit an AI system's behavior to reduce unwanted or unsafe outputs and actions.
Na way of using AI where person still dey check, guide, or approve important part of the work or the result.
An approach where a person reviews, guides, or approves important parts of an AI process or its results.
Na how dem dey protect and handle information about people well, including how dem collect, use, keep, and share am.
The protection and appropriate handling of information about people, including how it is collected, used, stored, and shared.
Na when people purposely test AI with difficult or attack-style input to find weakness and risk.
Deliberately testing an AI system with challenging or adversarial inputs to find weaknesses and risks.
Na when dem design and use AI with care for safety, fairness, privacy, reliability, clear explanation, and how e go affect people.
The design and use of AI with attention to safety, fairness, privacy, reliability, transparency, and its effects on people.
Na visible or hidden mark wey dem add to content to show where e come from, who own am, or say AI generate am.
A visible or hidden mark added to content to indicate its origin, ownership, or AI-generated status.