Build Practical AI Fluency for the Modern Workplace

Computer systems that perform tasks typically requiring human intelligence, including reasoning, learning, decision-making, and language understanding.

An AI model trained on massive text datasets that generates human-like text by predicting the most probable next words in a sequence.

AI systems that create new content (text, images, code, audio) rather than just analyzing or classifying existing data.

A subset of AI where systems improve their performance on tasks through experience and data, without being explicitly programmed for each scenario.

The branch of AI focused on enabling computers to understand, interpret, and generate human language.

A computing system inspired by biological neurons, consisting of layered nodes that process information and learn patterns from data.

The basic unit of text that LLMs process, roughly ¾ of a word in English. Models have token limits that constrain input and output length.

The neural network architecture behind modern LLMs (GPT, Claude, Gemini). Uses "attention" mechanisms to understand relationships between words regardless of distance.

AI systems that can autonomously plan, execute multi-step tasks, use external tools, and adjust their approach based on results.

A large AI model trained on broad data that can be adapted to many different tasks. The base for tools like ChatGPT, Claude, and Gemini.

The text input you give to an AI model: your instructions, questions, or context that shapes the response.

The practice of crafting effective prompts to get useful, accurate, and relevant outputs from AI models.

Background instructions that set the AI's role, behavior, and constraints before the user's conversation begins.

Including examples in your prompt to show the AI the pattern, format, or style you want it to follow.

Asking the AI to perform a task without providing examples, relying entirely on clear instructions.

A prompting technique that asks the AI to show its reasoning step by step, which often improves accuracy for complex tasks.

The maximum amount of text (measured in tokens) that a model can process in a single conversation, including both your input and its output.

A setting that controls randomness in AI output. Lower temperature = more predictable; higher = more creative and varied.

When an AI generates plausible-sounding but factually incorrect or fabricated information, including fake citations, invented statistics, or non-existent sources.

Systematic skew in AI outputs that reflects imbalances in training data, producing unfair, stereotyped, or unrepresentative results.

The protection of personal and sensitive information from unauthorized access, especially important when data is shared with AI services.

Any data that could identify a specific person, such as names, email addresses, phone numbers, social security numbers, and employee IDs.

When employees use personal AI accounts and non-approved tools for work tasks, creating security, compliance, and quality risks.

AI-generated content that is published or delivered without meaningful human review, editing, or quality control. Often generic, incorrect, or tone-deaf.

Technical and policy controls that prevent AI systems from producing harmful, off-topic, or non-compliant outputs.

The ability to understand AI capabilities and limitations, use AI tools effectively, maintain quality standards, and make informed decisions about when and how to apply AI at work.

An organization's or individual's progression from basic awareness of AI through experimental use to strategic, measured integration into workflows.

Using AI to enhance human capabilities (making people faster, more thorough, or more creative) rather than replacing human judgment.

A workflow design where AI handles initial processing but humans review, edit, and approve all outputs before they are used or shared.

How long it takes to produce an initial version of a deliverable. A key metric for measuring AI's impact on productivity.

Measuring the return on AI investment by comparing time saved, quality improvements, and error reduction against tool costs and learning time.