In the Substack app, click the play button at the top of a post to start narration. To adjust audio speed, tap the playback bar at the bottom of your screen.
As with most subjects in our daily lives, you don't need to memorize these terms. Simply keep this reference on hand to decode and evaluate incoming information.
Looking for a framework behind these terms? See: A Framework & Resource Portal for Exploring AI.
AI-GENERATED IMAGES — Media generated using text prompts, parameters, and generative fill tools to create or alter visual images. [How-To Geek]
AI-GENERATED TEXT — Text, sentences, or conversations produced by AI, often following predictable patterns, such as a steady tone, repetitive phrasing, and structured transitions. [Grammarly] It is “often not identified as such but presented as language written by humans, raising concerns about novel forms of deception and manipulation.” [PNAS]
ALGORITHM — A finite, well-defined set of rules, instructions, or steps designed to solve a problem, perform calculations, or train a model. [GeeksforGeeks]
ARTIFICIAL INTELLIGENCE (AI) — Technology that enables computers and machines to simulate human capabilities, particularly pattern recognition, speech recognition, translation, and prediction to perform tasks such as decision-making, self-correction, and content generation. [IBM, IBM, Zendesk]
BIAS (in AI) — Systematic errors in outputs resulting from prejudiced training data or flawed algorithms, leading to inaccurate or improper outcomes. [Columbia Business School]
CHATBOT — A software program designed to simulate human conversation. [IBM]
COMPUTER VISION — A subfield of AI that equips machines to process, analyze, and interpret visual input (images and video) to derive meaningful information. [IBM]
CONTEXT WINDOW — The maximum amount of text or data (measured in tokens) an AI model can evaluate and hold in memory at one time. [Columbia Business School]
DATA MINING — The process of discovering patterns, anomalies, and correlations from large datasets, across databases and digital repositories. [University of Illinois Springfield]
DATASET — A structured collection of data used to train or evaluate a model.
DEEP LEARNING (DL) — An advanced subset of machine learning designed to continuously analyze massive datasets and build complex knowledge representations over time. [Zendesk] The term may be loosely used synonymously with machine learning since it functions in a similar way — “however, its capabilities are different.” [University of Illinois Springfield]
DEEPFAKE — Synthetic media (images, video, or audio) generated by AI that convincingly portrays real people saying or doing things that never occurred in reality. [Britannica]
GENERAL AI (AGI) — In contrast to Narrow AI, AGI is “capable of performing a wide range of tasks across various domains with adaptability and reasoning.” [MIT]
GENERATIVE AI / GEN AI — AI models capable of creating original content, such as text, images, video, audio, or software code, in response to a prompt [IBM]
GUARDRAILS (in AI) — Technical controls, policies, and safety constraints built into AI systems to intercept, block, and mitigate problematic or unsafe outputs in real time. [GeeksforGeeks]
HALLUCINATION (in AI) — Instances where an AI model generates false, unsupported, or entirely fabricated information presented confidently as fact. It “may take the form of fake facts, invented quotes, incorrect citations, or completely fabricated people, places, or events.” [Science News Today]
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.