Artificial Intelligence Fundamentals
Artificial intelligence is the field of building systems that perform tasks normally requiring human intelligence: recognizing patterns, making decisions, understanding language, and learning from data. It spans rule-based systems, machine learning, deep learning, and the engineering needed to deploy them.
itArtificial intelligence and machine learning | OpenSkills.info
Intro
Artificial Intelligence Fundamentals
Artificial intelligence is not one algorithm, one product, or one kind of output. It is a broad family of machine-based systems. These systems take input and infer how to produce predictions, content, recommendations, or decisions for an objective.
That definition gives you a useful boundary. An AI system can recommend a song, classify an image, forecast demand, generate text, or choose an action. The output may affect software, a business process, or a physical device. The amount of autonomy varies by system.
The practical question is not, "Does this look intelligent?" Ask what objective the system serves, what input it receives, what output it produces, and what that output influences.
The system is larger than the model
An AI model maps input to output. An AI system surrounds one or more models with data collection, application logic, interfaces, monitoring, people, and operating controls.
This distinction matters. A model can score well in isolation while the deployed system fails its users. Input may differ from training data. An interface may hide uncertainty. A process may give an output more authority than it deserves. You evaluate the complete system in its intended context, not only the model.
Use this five-part mental model:
- Objective — the result the system is designed to pursue.
- Input — data, rules, prompts, sensor readings, or other signals.
- Model — the learned or hand-built representation used for inference.
- Output — a prediction, content, recommendation, or decision.
- Context — the people, process, environment, and consequences around the output.
Continue the course
This section is part of the paid course.
See pricing to subscribe, or log in if you already have access.
Where this skill leads
Relevant careers
See how this topic contributes to broader role-level skill maps.
Sources
- https://www.oecd.org/content/dam/oecd/en/publications/reports/2024/03/explanatory-memorandum-on-the-updated-oecd-definition-of-an-ai-system_3c815e51/623da898-en.pdf
Supports
- The updated AI-system definition based on objectives, input, inference, outputs, physical or virtual environments, autonomy, and adaptiveness
- Predictions, content, recommendations, and decisions as output categories
- Machine learning and knowledge-based approaches as categories encompassed by AI
- Computer vision, natural language processing, speech recognition, decision support, and robotics as application areas
- Human roles and supervision across design, data collection, development, validation, deployment, operation, and monitoring
- Training as performance improvement through exposure to data, and inference as generating output from input
- An AI model as a core component of an AI system and the distinction between models and complete systems
- Statistical models, decision trees, and neural networks as examples of AI models
- Supervised, unsupervised, and reinforcement learning as ways to build models
- Recognition, event detection, forecasting, personalization, interaction support, optimization, and knowledge-based reasoning as representative tasks
- https://developers.google.com/machine-learning/intro-to-ml/what-is-ml
Supports
- Machine learning as training a model from data to make predictions or generate content
- Supervised, unsupervised, reinforcement, and generative AI learning categories
- Labeled examples for supervised learning and regression versus classification
- Clustering as grouping similar data without predefined categories
- Reinforcement learning rewards, penalties, and policies
- Generative AI content outputs and the use of further supervised or reinforcement training
- https://developers.google.com/machine-learning/intro-to-ml/supervised
Supports
- Data, model, training, evaluation, and inference as core supervised learning concepts
- Features, labels, labeled examples, and unlabeled examples
- Evaluation on unseen data before inference on new input
- Dataset size and diversity as contributors to performance on new data
- https://developers.google.com/machine-learning/crash-course/
Supports
- The course progression through regression, classification, data preparation, generalization, overfitting, neural networks, embeddings, large language models, production systems, and fairness
- Neural network architectures with perceptrons, hidden layers, and activation functions
- Production concerns including training, inference, deployment testing, and monitoring
- https://developers.google.com/machine-learning/glossary/fundamentals
Supports
- Definitions for model, feature, label, inference, generalization, and overfitting
- A neural network as a model with hidden layers
- A deep model or deep neural network as a neural network with more than one hidden layer
- https://developers.google.com/machine-learning/crash-course/llm
Supports
- A language model as estimating probabilities for tokens or token sequences in context
- Tokens as words, subwords, or characters used as language-modeling units
- Text generation, translation, and summarization as language-model tasks
- https://airc.nist.gov/airmf-resources/airmf/
Supports
- The AI Risk Management Framework as voluntary guidance across design, development, use, and evaluation
- AI actors and risk management across the AI lifecycle
- Govern, Map, Measure, and Manage as the four AI RMF Core functions
- Profiles as contextual implementations based on use, risk tolerance, and resources
- https://airc.nist.gov/airmf-resources/airmf/3-sec-characteristics/
Supports
- Valid and reliable, safe, secure and resilient, accountable and transparent, explainable and interpretable, privacy-enhanced, and fair with harmful bias managed as trustworthy AI characteristics
- Evaluation with realistic test sets representative of expected conditions
- Ongoing testing and monitoring for validity and reliability
- Context-dependent tradeoffs among trustworthiness characteristics
- Transparency, explainability, and interpretability as distinct and related characteristics
- Systemic, computational and statistical, and human-cognitive bias categories
- Contextual assessment of whether AI is appropriate based on risks, impacts, costs, and benefits
- Adversarial examples and data poisoning as AI security concerns