AI fundamentals

  • 6 minutes

In simple terms, artificial intelligence (AI) is the design and study of machines that perform tasks that required human brainpower in the past. AI encompasses many aspects of intelligence, such as reasoning, decision-making, learning from mistakes, communication, and problem-solving. This doesn't diminish the need for critical thinking and human oversight across the application and utilization of AI tools.

Core AI terminology and concepts

Artificial intelligence (AI): Technology that enables computers to perform tasks that normally require human intelligence, such as analyzing information, recognizing patterns, and making suggestions. It's where machine learning (ML), natural language processing (NLP), large language models (LLM), and other techniques work together to help computers learn, understand, and generate outputs in ways that support human tasks.

Machine learning (ML): A branch of AI involving computer algorithms that can ‘learn' by recognizing patterns in data and use those patterns to make future predictions or provide other useful outputs.

Machine learning includes:

  • Supervised learning: The machine is given a set of labeled training data, meaning the input data is paired with the correct output. The model can learn to predict labels or classify new, unlabeled data based on this training.
  • Unsupervised learning: Training a model on a dataset without labeled outcomes. The AI algorithm identifies patterns, structures, or relationships in the data without predefined labels.
  • Reinforcement learning: A type of machine learning where an AI system learns by interacting with an environment and improving its actions over time based on feedback in the form of rewards or penalties, essentially learning through trial and error to achieve goals without explicit programming. It mimics human learning, allowing AI to make sequential decisions in changing situations, finding the best strategies over time.
  • Deep learning (DL): A subset of machine learning that uses multi-layered neural networks (deep neural networks) to learn complex patterns in data, enabling systems to make predictions or decisions from large volumes of unstructured and unlabeled data.
  • Artificial neural networks (ANN): Computing systems inspired by the human brain, consisting of interconnected layers of artificial neurons (nodes) trained to recognize patterns from data in order to map inputs to outputs.
  • Large language models (LLMs): A subset of generative AI (GenAI) focused specifically on understanding and generating human language (text).
  • Natural language processing (NLP): An AI capability that enables computers to understand, interpret, translate and generate human language.
  • Prompts: A prompt interface is the user-facing mechanism through which people interact with a GenAI system by entering, structuring, and refining prompts. It provides the space and tools—such as text fields, templates, system instructions, parameters such as retrieval augmented generation (RAG), or examples—that help users clearly communicate their intent and guide the AI's outputs. In simple terms, the prompt interface is where users "talk to" the AI and shape how it responds.
  • Retrieval augmented generation (RAG): This is a two-step technique that combines retrieval of relevant information from trusted data sources (such as documents, databases or knowledge repositories) and uses that information in response generation, resulting in improved accuracy and reliability of GenAI tools.

How AI works

AI functions through the combined power of three major components: machine learning models, natural language processing, and data. Each plays a distinct but interconnected role in how AI systems learn, interpret, and generate outputs.

How the components interconnect

AI's effectiveness stems from how these components work together: Deep learning models combine ML, NLP, and large datasets to create powerful tools such as language models, image classifiers, and speech-to-text systems.

  • Data provides the raw material.
  • ML learns patterns in that data, whether numbers, images, or past decisions.
  • NLP enables the system to understand prompts and generate human language, allowing individuals to interact with AI in natural text or speech rather than code.

Together, this interconnected ecosystem allows AI to support nonprofits with tasks such as analyzing large datasets, drafting reports, translating content, detecting trends, and providing faster insights for programs, operations, and decision-making.