AI Basic Knowledge
What is AI?
A.I. stands for Artificial Intelligence. AI systems use hardware, software, algorithms, and data to approximate human intelligence. Such systems are usually designed to accomplish specific tasks (e.g. make decisions, discover patterns, and perform functions). AI is currently used in many areas of technology, from self-driving cars and virtual assistants like Siri and Alexa, to spelling and grammar checking software in Microsoft Word.
There are different approaches to building AI systems. Two of the primary approaches are:
Rule-based systems, which use explicit and predefined rules provided by human programmers.
Machine Learning systems, which use statistics and algorithms to approximate "learning," so that the system can be more versatile and adaptable, without needing an explicit rule for each application. Many newer AI systems use machine learning. Recently there have been some exciting developments in “generative AI”: artificial intelligence systems which are used to generate text, computer code, images, music, and many other forms of media.
Definitions
Artificial Intelligence (AI): The simulation of human intelligence processes by machines, typically involving learning, reasoning, problem-solving, perception, and decision-making.
Data/Input: Information used by AI systems to learn and make decisions, typically represented in numerical or categorical form and obtained from various sources such as sensors, databases, or manual input.
Output/Prediction: The result or outcome generated by AI systems based on input data and learned patterns, which may include classifications, predictions, recommendations, or other forms of decision-making.
Generalized AI: Has human-like intellect and the capacity to comprehend, learn, and apply information across a wide variety of areas and tasks. The type of AI that appears in popular media, but does not yet exist.
Narrow AI: Focused on narrow applications and needs to be trained to perform a specific task(s). All current AI systems fall under this category.
Generative AI:– Generative AI is a branch of artificial intelligence that uses machine-learning systems. These are built by humans using large amounts of data. The data is used to build a model that maps vast arrays of statistical relationships within the data (patterns). Once the model has been set up and is running, it performs its functions without human supervision.
For example, ChatGPT uses a language model built from text scraped from the internet that has been processed to produce a kind of map of probable meaning and probable relationships between meanings. This model (known as a “Large Language Model” or LLM) can then be used to generate text responses to user prompts.
Algorithmic Bias - video
AI learning assistants
AI assistive technology
AI-driven tools, devices, and software that support people with disabilities in navigating otherwise inaccessible spaces or content. Speech recognition and computer vision are among the most popular categories of AI-powered assistive tools.
Examples: SeeingAI, EasyReader, Live Caption, VoiceIn, High Contrast, Read Aloud
Chatbots
Chatbots generate written responses, answer questions, and check work for spelling, grammar, etc.
Examples: ChatGPT, Perplexity AI, Google Bard, Bing AI, Jasper AI, Gemini, example with visual
Writing assistants
AI writing assistants edit, rephrase, and rewrite text to improve writing.
Examples: Grammarly, QuillBot, Hemingway Editor
Homework apps
Homework help sites use AI to generate study guides and explain steps to solve a problem. Homework sites and apps are trained to provide the information needed to arrive at answers. This is slightly different than AI chatbots: those are trained to give answers without "showing the work" (unless specifically prompted).
Examples: Course Hero, Photomath, Wolfram Alpha, Socratic
Voice assistants
Voice assistants take spoken commands to provide answers to questions and operate portions of devices, such as search engines and music apps.
Examples: Siri, Alexa, Google Assistant, Cortana