An Artificial Intelligence (AI) system [is] a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments.
Organisation for Economic Co-operation and Development (OECD)
Seven use cases for AI, that can run in parallel with one another:
- Hyper-personalization - displaying relevant content in online searching; targeted marketing (e.g. Netflix suggested shows)
-
Conversation and human interaction - e.g. chatbots for mental health care; Alexa, Siri, & Google Assistant
-
Pattern and anomaly detection - detecting fraud; flagging purchases in unusual amounts or locations
-
Recognition - facial/voice recognition; e.g. combine with anomaly detection to review photos of skin and provide potential diagnoses
-
Goal driven systems - finding optimal solutions to specific problems; gaming (e.g. Deep Blue, the computer that beat the world chess champion in 1997)
-
Predictive analytics and decision support - prediction about human behavior; weather forecasting
-
Autonomous systems - able to interact with surroundings and achieve objectives with little to no human involvement (e.g. autonomous vehicles)
From Artificial Intelligence & Responsible Business Conduct, OECD
Large Language Models (LLM)
- Very large "deep learning" AI models that are pre-trained using immense amounts of data
- Incredibly flexible, can be used to consider billions of parameters to:
- answer questions
- complete sentences
- summarize documents
- translate languages
- Demonstrate an ability to make predictions based on a small number of input/prompts
- Can be used for generative AI to produce content based on human input/prompts
From What are Large Language Models (LLM)?, AWS
