Core AI concepts
The foundation layer of the AIGP vocabulary: what AI is, what it runs on, and its classic forms. Artificial intelligence is defined as machine-based systems that infer from inputs how to generate outputs.
This first cluster is the bedrock of the AIGP vocabulary: what AI is, what it runs on, and its classic forms. Artificial intelligence itself is defined as machine-based systems that, for given objectives, infer from inputs how to generate outputs - predictions, content, recommendations or decisions - that influence real or virtual environments.
- Capability spectrum: an Algorithm is a fixed recipe; Autonomy is how far a system acts without human intervention; Agentic AI sits at the high-autonomy end, chaining tools across end-to-end workflows.
- Hypothetical horizon: Artificial general intelligence (AGI) would match or exceed human cognition across any task - still hypothetical, unlike today's narrow AI.
- Heritage forms: the Expert system (knowledge base + if-then inference engine), the Turing test, and Robotics predate modern ML.
- Application fields: Natural language processing (NLP), Computer vision and the Chatbot are the classic AI application areas.
- Compute - the GPUs and processing power - is both a cost driver and a policy lever for regulating frontier models.
The exam reuses the AI definition almost verbatim: AI infers from inputs how to generate outputs (predictions, content, recommendations or decisions) that influence real or virtual environments. Note it says infer, not just compute.