AI is a hot topic for financial innovators, but few can keep up with the technical terms. Don’t drown in the alphabet soup of jargon. Here are the most used AI terms in plain English.
AI: Artificial Intelligence. A computer that can understand natural language and develop learning skills. More sophisticated than automation, AI enables computers to make complex, judgment-driven decisions that mirror the human process, and then learn from the outcome of those decisions.
Analogical Reasoning: The ability to solve a new problem by comparing the outcomes of past experiences. Humans do it all of the time, computers are just learning.
ANN: Artificial Neural Networks. Machine learning models that seek to imitate the structure of the brain’s neural network.
API: Application programming interface. A set of protocols that allows two (or more) systems to interact with each other. When you choose “log in with Facebook,” on a third party website, an API to connects the two systems. Internal APIs enable connectivity across financial institutions. External APIs connect third-party developers to those institutions’ systems.
Chatbot: A computer program that simulates human conversation through artificial intelligence. Some predict the chatbot will replace the bank teller.
Deep Learning: A subset of machine learning that uses multiple processing layers to make decisions.
NLP: Natural language processing. A computer’s ability to derive meaning from human language through machine interpretation of text and speech recognition.
Robo Advisor: Wealth Management of the Future. A digital financial advisor that assesses your goals and risk tolerance to build an algorithmically managed investment portfolio with little or no human interaction. Soon to be powered by AI.
SDK: Software development kit. A set of development tools that makes it easy to implement an API. SDKs can contain pre-built screens, or a designated workflow to help developers communicate with APIs.
Weak AI: Artificial intelligence that uses few layers of processing, and focuses on a small set of tasks. Most AI in use today is weak AI.