| Artificial Intelligence (AI) |
Computer systems designed to mimic human tasks like understanding language, making decisions, and learning from experience. |
Siri, Alexa, self-driving cars |
| Machine Learning |
A type of AI that learns from data to recognize patterns and improve performance over time without being explicitly programmed. |
Spam filters, Netflix recommendations |
| Large Language Models (LLMs) |
AI models trained on vast amounts of text data to generate human-like responses, such as GPT or Bing Chat. |
ChatGPT, Bing Chat, Gemini |
| Generative AI |
AI that creates new content—text, images, music, code—based on learned patterns rather than repeating known material. |
DALL-E, Copilot, Stable Diffusion |
| Hallucinations |
When AI produces false or made-up information with confidence; an ongoing challenge in generative AI. |
AI says "Abraham Lincoln was born in 1945" |
| Responsible AI |
Principles and practices to ensure AI is ethical, safe, inclusive, fair, and transparent across its design and deployment. |
Microsoft Responsible AI Standard |
| Multimodal Models |
AI systems that understand and combine different input types like text, images, and sound for more complex interactions. |
GPT-4 Vision, Gemini, Copilot Studio |
| Prompts |
Instructions or inputs (text, image, code) used to guide AI systems in generating desired outputs. |
"Write a poem about summer", uploading a photo for description |
| Copilots |
AI tools integrated into software (e.g. Microsoft 365) that assist users by offering suggestions, automating tasks, and enhancing productivity. |
Microsoft 365 Copilot, GitHub Copilot |
| Reasoning / Planning |
AI’s ability to break down complex goals into logical steps to plan or solve problems—like planning a trip or building a to-do list. |
Copilot planning a meeting, AI generating a travel itinerary |