What is AI? A broad field, not one tool
Artificial intelligence is the broad field of building computer systems that perform tasks associated with intelligence, such as recognizing patterns, making predictions, understanding language or choosing actions.
No ranking, interview or hiring outcome is guaranteed.
A simple map
- AIThe broad field and the systems built within it.
- Machine learningAn approach in which models learn useful patterns from data or experience.
- Deep learningMachine learning based on multi-layer neural networks.
- Large language modelsDeep-learning models trained to work with sequences of language tokens.
- AI productsApplications that combine models with interfaces, data, rules, tools and human workflows.
Capability is not certainty
A language model predicts useful text from context; fluent output can still be false. A predictive model can work on evaluation data and fail after the population changes. Responsible use requires testing the specific system on the specific task.
That is why consequential claims, applications and decisions need source checks and human review rather than trust in the label AI.
A system is more than its model
An AI product usually surrounds a model with data collection, instructions, retrieval, software rules, interfaces and monitoring. Two products using a similar model can behave differently because they supply different context, permit different actions and check outputs in different ways. Evaluate the complete workflow rather than treating the model name as a specification.
The right test depends on the task. A classifier needs representative labeled examples and meaningful error costs. A writing assistant needs factual and source checks. A tool-using system needs permission boundaries and action logs. Broad claims that a system is intelligent say less than a concrete account of its inputs, outputs, evaluation and failure handling.
Questions to ask of an AI system
- What evidence enters?Identify training assumptions where known, the information supplied at use time and anything important the system cannot observe.
- What can it change?Separate suggestions from actions and find where a person reviews consequential output before it affects the outside world.
- How was it tested?Look for evaluation on the actual task, important groups and failure cases rather than relying on a general benchmark or fluent example.
Questions
Is AI the same as ChatGPT?
No. AI is the broad field; ChatGPT is one product built with language models.
Is all AI machine learning?
No. Machine learning is a major approach within AI, but the broader field also includes rule-based and search methods.
Is every rule-based computer program AI?
No single boundary satisfies every context. In practice, explain the specific technique and capability rather than relying on the broad label alone.
Why can an AI system fail even when its model is capable?
Poor data, missing context, weak instructions, unsafe tools, bad evaluation or an unsuitable human workflow can make the complete system unreliable.
Sources
Facts about other courses were read from each provider's own pages on the dates shown. Prices and free limits change, so check the provider's page before you enrol.