On November 30, 2022, OpenAI released ChatGPT, a program that revolutionized how humans interact with technology and the world around them. Unlike most technological tools that came before it, ChatGPT could respond to almost any question and independently generate text. Since ChatGPT was released, companies like Microsoft, X (formerly Twitter), Google, and IBM have also released similar tools and integrated them into current product offerings.
What is AI?
Although there is no universally agreed-upon definition as to what artificial intelligence is, in simple terms, it refers to a machine’s ability “to perform the cognitive functions we associate with human minds, such as perceiving, reasoning, learning, interacting with an environment, problem-solving, and even exercising creativity.” While AI is considered a recent invention or emerging technology, it has been a part of society for some time, with tools like predictive text, voice assistants, and chatbots all used almost daily.
While AI appears in many different forms, the two most common are generative AI and machine learning. Each form has its distinct functions and use cases. The most well-known form of AI is generative artificial intelligence, known for its ability to generate content like text, audio, video, or images from human prompts. Generative AI relies on Large Language Models (LLMs) to collect vast amounts of text online from online sources that can then be trained to respond to specific prompts from a human operator. Given its ability to generate content and training on internet sources, generative AI could easily be used to draft responses, summarize texts, and quickly conduct research.
Unlike generative AI, Machine Learning (ML) takes vast amounts of data on a range of topics and uses them to make predictions. These predictions range from providing consumers with streaming recommendations and image recognition to predicting an individual’s likelihood of suffering from a disease or illness. ML is further categorized as supervised or unsupervised. Supervised ML requires labeled data sets to train algorithms to grow more accurate over time. For example, an algorithm trained with labeled images of cars can eventually learn to identify types of vehicles without human guidance. Unsupervised learning is used to find patterns and trends, which can then be used to predict future behavior
Read “Bringing Government Into the 21st Century: Artificial Intelligence and State Government Operations” here.
