Artificial Intelligence Terms You Need to Know!
David Savoie, CEO, LaCorp
Artificial Intelligence (AI) is one of the most talked-about topics among credit unions today in both the technology and business arenas. AI-based technology is impacting nearly all industries, and credit unions are no exception. Most experts feel we are at the earliest stage of AI-driven innovations in business.
Whether your credit union has adopted or plans to adopt AI enhancements in the near term, everyone in a management or public-facing position in credit unions should have a working knowledge of basic AI terminology.
The Credit Union Connection has partnered with igniteFI to host a unique webinar series, The Fintech Adventure! Our first virtual event will tackle a deep dive into AI for an informational hour you can’t miss on Sept. 19. Then, on Sept. 20, follow it up with speed dating sessions with the Credit Union Business Partners in the trenches. Space is limited, so learn more and register for Day 1 here and Day 2 here!
Whether you are new to AI or have some expertise and would like a refresher, we think you will find that taking a few minutes today to go over some basic AI terms will benefit you greatly in the coming days.
To start, Artificial Intelligence (AI) can be defined as “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.”
AI used in business can be divided into three main types: Generative AI is “an advanced technological approach that enables the creation of content including text, images, and videos.” ChatGPT is an example of a generative AI system used for generating text like marketing materials and business letters. While generative AI systems may seem human in nature, they work by analyzing vast amounts of sample data to predict what text is likely to follow the user’s prompt and do not possess human consciousness or emotions themselves.
Natural Language Processing (NLP) is a “subfield of AI and computational linguistics that focuses on enabling machines to understand, interpret, and generate human language to be understood by humans.” Amazon’s Alexa, Apple’s Siri, and Google Translate are examples of NLP systems.
Machine Learning is “a branch of artificial intelligence (AI) and computer science that focuses on using data and algorithms to enable AI to imitate the way that humans learn, gradually improving its accuracy.” Neural Networks, a subfield of Machine Learning, are “systems modeled after the human brain and are mathematical systems that actively learn skills by identifying and analyzing statistical patterns in data.” Reinforcement Learning is a type of Machine Learning in which models learn optimal decision-making strategies through feedback cycles, with human interaction playing a pivotal role in refining the learning process. Cash Ordering and Inventory systems that adjust their order frequencies and sizes based on past and ongoing order history are examples of Machine Learning-based systems.
Human-In-The-Loop (HILT) is an element of training a Machine Learning model, defined as “an iterative process whereby a human (or a team) interacts with an algorithmically generated system.” An example would be a cash ordering system where human management gives final approval to ordering decisions made by the model and the model uses these decisions as inputs to further refine its future recommendations. HILT can also refer to humans, such as data scientists, responsible for identifying and labeling historical real-world data used to train the model.
Robotic Process Automation (RPA), also known as software robotics, are “systems that use intelligent automation technologies to perform repetitive office tasks of human workers, such as extracting data, filling in forms, and moving files.” RPA can allow humans to focus on more exciting and meaningful tasks.
Before we close, here are two basic caveats to keep in mind about AI systems:
Hallucinations are “occurrences where large language models generate factually inaccurate or illogical answers due to data and architecture constraints.” Because generative AI writes text based on databases of existing text and is not consciously aware of actual reality, it can generate false statements or “hallucinations,” such as events that have not occurred or quoting statements not actually attributable to the person being quoted. Accordingly, most text generated by Generative AI models, such as business letters and emails, need human review before final approval.
PII and proprietary data: When using publicly available AI models like ChatGPT or Claude, credit union users should be cautioned not to enter any member’s actual Personally Identifiable Information or any credit union proprietary information since any data entered into the system becomes the property of the company that owns the system and is subject to inadvertent exposure.
As AI continues to evolve and integrate with the systems we use each day, we hope these basic terms will equip you to continue your journey of AI learning!
