We generated these definitions by uploading the content of our AI Ethics Learning Toolkit into Google’s NotebookLM and then creating a glossary.

---

AI Divide / Digital Divide 

The growing gap between individuals, communities, and companies who have access to, use of, and skills related to cutting-edge Artificial Intelligence technologies.

AI Hallucination

An instance where an AI model generates misleading, inaccurate, or entirely fabricated content, often without a clear basis in its training data. Many favor alternative terms (bullshit, machine errors, AI mirage), so as not to assign human qualities to AI technology.

AI Hype / AI Boosterism

An outsized enthusiasm for the benefits of Artificial Intelligence by individuals, organizations, and/or the media, with little acknowledgement of the harms these technologies may cause. 

AI Literacy

The ability to understand how Artificial Intelligence (AI) works, recognize its strengths and limitations, critically evaluate AI tools and outputs, and thoughtfully consider the ethical, social, and practical implications of its use. 

AI Privilege / Information Privilege

Related to the AI Divide, AI privilege refers to the benefits of one’s access to, use of, and skills related to cutting-edge Artificial Intelligence technologies. 

AI Slop

The flood of low-quality, sometimes misleading, AI-generated content (text, images, video) present on the internet, social media, and other information/publication platforms. 

Bias (data)

Bias that is present in the training and fine-tuning data of AI models. Bias negatively affects the models functions and outputs. 

Content Moderation

The process (human, or automated) of reviewing and filtering online content to remove material that violates platform guidelines or is deemed inappropriate or harmful. 

Copyright Infringement

The reproduction, distribution, performance, or derivative of a copyrighted work without the permission of the copyright owner. 

Critical Thinking

The analysis and evaluation of an issue in order to form a judgment. In the context of AI, it involves the skills of evaluating AI-generated information and questioning its assumptions. 

Data Labeling / Data Annotation

The process of annotating raw data with meaningful tags or labels to provide context and structure for machine learning models.

Deepfakes

Synthetic media (images, audio, or video) that have been altered or created using AI to convincingly depict someone saying or doing something they did not actually say or do. These may be somewhat benign celebrity deepfakes (Taylor swift peddling Le Creuset cookware), or dangerously misleading (ex. Zelensky deepfakes).

Fair Use

A legal doctrine that permits limited use of copyrighted material without permission from the copyright holder for purposes such as criticism, news reporting, teaching, scholarship, or research. 

Garbage in, Garbage out (GIGO)

The phenomenon where flawed, biased, or low-quality data or input leads to outputs that are similarly flawed, biased, or of poor quality. 

Ghost Work / Ghost Labor

The often invisible human workers who make AI systems function by working on tasks like data labeling, content moderation, and transcription. 

Mis/Disinformation

(Dis)information – False or inaccurate information that is deliberately created and spread with the intent to deceive or cause harm. (Mis)information has the same effects but is not intentional.

Model Training

The process of using data to teach a machine learning model to identify patterns and make predictions or decisions. 

Monoculture

In the context of AI training data, it refers to the overrepresentation of certain viewpoints, languages, or perspectives (often from the Global North and dominant cultures) leading to a lack of diversity in the AI’s understanding and outputs. 

Overreliance

Excessive dependence on AI tools or outputs, potentially leading to a decline in human skills like critical thinking and independent decision making. 

Privacy

The right of individuals to control the collection, use, and dissemination of their personal information. In AI, it relates to how the user data and information used for training models are handled. 

Trust

Confidence in the reliability, truth or ability of something. The rapid spread of AI-generated content, particularly hallucinations and AI slop, muddies the waters of what is reliable information online.