1.1 What Is Artificial Intelligence?
Key Takeaways
- Artificial intelligence (AI) is software that imitates human behaviors and capabilities — including visual perception, speech recognition, decision-making, and language understanding.
- AI is an umbrella term that encompasses machine learning, deep learning, computer vision, natural language processing, and generative AI.
- Machine learning is a subset of AI where models learn patterns from data; deep learning is a subset of ML using multi-layered neural networks.
- AI workloads on Azure include prediction, anomaly detection, computer vision, NLP, conversational AI, and generative AI.
- The AI-900 exam tests your ability to identify which type of AI workload applies to a given business scenario.
Quick Answer: Artificial intelligence (AI) is software that imitates human capabilities such as visual perception, speech recognition, decision-making, and language understanding. AI encompasses machine learning, deep learning, computer vision, NLP, and generative AI. On Azure, these capabilities are delivered through services like Azure AI Vision, Azure AI Language, Azure AI Speech, and Azure OpenAI Service.
Defining AI
At its core, artificial intelligence refers to computer systems that can perform tasks that typically require human intelligence. This includes:
- Visual perception — recognizing objects, faces, and text in images
- Speech recognition — converting spoken words into text
- Decision-making — making predictions or recommendations based on data
- Language understanding — comprehending and generating human language
- Content generation — creating new text, images, code, or other content
AI is not a single technology. It is an umbrella term that covers many approaches, techniques, and specializations:
The AI Hierarchy
Understanding how AI concepts nest within each other is critical for the AI-900:
Artificial Intelligence (broadest)
├── Machine Learning (learning from data)
│ ├── Supervised Learning (labeled data)
│ │ ├── Regression (predict numbers)
│ │ └── Classification (predict categories)
│ ├── Unsupervised Learning (no labels)
│ │ └── Clustering (find groups)
│ └── Deep Learning (neural networks)
│ ├── Computer Vision
│ ├── Natural Language Processing
│ └── Generative AI
└── Rule-Based Systems (explicit rules, no learning)
| Concept | Definition | Example |
|---|---|---|
| Artificial Intelligence | Software that imitates human capabilities | A virtual assistant that answers questions |
| Machine Learning | AI that learns from data without explicit programming | A spam filter that improves over time |
| Deep Learning | ML using multi-layered neural networks | Image recognition that identifies dog breeds |
| Computer Vision | AI that processes and understands visual information | Reading text from a photograph (OCR) |
| Natural Language Processing | AI that understands human language | Sentiment analysis of customer reviews |
| Generative AI | AI that creates new original content | ChatGPT generating a summary of an article |
On the Exam: Questions often present a scenario and ask you to identify the type of AI workload. The key is recognizing the INPUT and OUTPUT: images = computer vision, text understanding = NLP, new content creation = generative AI, numerical predictions = regression, category predictions = classification.
Common AI Workloads
Microsoft organizes AI capabilities into several workload types that you must recognize on the exam:
1. Prediction and Forecasting
Using historical data to predict future outcomes. Examples include sales forecasting, stock price prediction, and weather modeling. This typically involves regression machine learning models.
2. Classification
Assigning items to categories based on their features. Examples include spam detection (spam or not spam), medical diagnosis (disease or no disease), and credit risk assessment (approve or deny). This uses classification machine learning models.
3. Anomaly Detection
Identifying unusual patterns that do not conform to expected behavior. Examples include fraud detection, manufacturing quality control, and network intrusion detection. Azure provides the Anomaly Detector service for this workload.
4. Computer Vision
Extracting meaningful information from images and videos. This includes image classification, object detection, facial recognition, and optical character recognition (OCR). Azure provides Azure AI Vision, Custom Vision, and Face API services.
5. Natural Language Processing (NLP)
Understanding, interpreting, and generating human language. This includes sentiment analysis, entity extraction, translation, and speech recognition. Azure provides Azure AI Language, Azure AI Speech, and Azure AI Translator services.
6. Conversational AI
Building AI-powered chatbots and virtual assistants that can hold natural conversations. Azure provides Azure AI Bot Service and Azure OpenAI Service for this workload.
7. Generative AI
Creating new content such as text, images, code, and audio. This uses large language models (LLMs) like GPT-4o and image generation models. Azure provides Azure OpenAI Service for this workload.
8. Knowledge Mining
Extracting insights from large volumes of unstructured data (documents, images, audio). Azure provides Azure AI Search with cognitive skills enrichment for this workload.
AI on Azure — Service Landscape
| AI Workload | Azure Service | What It Does |
|---|---|---|
| Image analysis | Azure AI Vision | Analyze images — detect objects, read text, describe scenes |
| Custom image models | Azure AI Custom Vision | Train your own image classification or object detection models |
| Face detection | Azure AI Face | Detect, identify, and verify human faces |
| Text analysis | Azure AI Language | Sentiment analysis, entity recognition, key phrase extraction |
| Speech processing | Azure AI Speech | Speech-to-text, text-to-speech, speech translation |
| Translation | Azure AI Translator | Translate text between 100+ languages |
| Document processing | Azure AI Document Intelligence | Extract data from forms, invoices, receipts |
| Search and mining | Azure AI Search | Full-text and vector search with AI enrichment |
| Content safety | Azure AI Content Safety | Detect harmful content in text and images |
| Generative AI | Azure OpenAI Service | GPT models for text, DALL-E for images, Whisper for speech |
| Anomaly detection | Azure AI Anomaly Detector | Detect anomalies in time-series data |
| Bot framework | Azure AI Bot Service | Build conversational AI bots |
On the Exam: You do not need to know implementation details of these services for the AI-900. You need to know WHAT each service does and WHEN to use it. The AI-102 tests implementation; the AI-900 tests awareness and selection.
Which of the following best describes the relationship between AI, machine learning, and deep learning?
A company wants to automatically categorize customer support emails as "billing", "technical", or "general inquiry". Which type of AI workload is this?
Which Azure service would you use to extract text from a scanned document image?
Match each AI workload to the correct Azure service:
Match each item on the left with the correct item on the right
Azure AI Foundry: The Unified Platform
Microsoft has consolidated its AI tooling under Azure AI Foundry (formerly Azure AI Studio), a single environment for discovering, building, and deploying AI solutions. For the AI-900 you should recognize a few umbrella terms and how they relate:
| Term | What It Means |
|---|---|
| Azure AI services | The rebranded family of pre-built AI APIs — Vision, Language, Speech, Translator, Document Intelligence, Content Safety (formerly "Cognitive Services") |
| Azure AI Foundry | The portal and platform for building generative-AI apps, including a model catalog of foundation models |
| Azure Machine Learning | The platform for training, deploying, and managing your own custom ML models |
| Azure OpenAI Service | Access to OpenAI models (GPT, DALL-E, Whisper) hosted on Azure with enterprise security |
On the Exam: The exam was updated to use current names. "Cognitive Services" is now "Azure AI services," and generative-AI questions reference Azure AI Foundry and its model catalog. If you studied older material, mentally map the legacy names to the new ones.
Narrow AI vs. General AI
Every AI system on Azure today is narrow AI — software built to perform a specific task (read text, detect objects, generate a summary). General AI (a hypothetical system that matches human intelligence across any task) does not exist commercially and is not something Azure services provide. The AI-900 stays grounded in narrow, task-specific AI: knowing which workload and which service solves a given business problem.