GENERATIVE AI:
WHAT IT IS / HOW IT WORKS
Embedded Files
Generative artificial intelligence (AI) refers to a type of machine learning, whereby a computer can create content based on its ability to assess similar content. The generated content can take many forms, such as music, art, or writing. According to Baidoo-Anu & Anash (2023), generative modeling AI is “…an unsupervised or partially supervised machine learning framework which generates manmade relics via the use of statistics, probabilities etc. Through advances in deep learning (DL), the generative AI create artificial relics using existing digital content such as but not limited to video, images/graphics, text, audio, video by examining training examples; learning their patterns and distribution” (p. 3).
Essentially, the program must be provided examples of the type of content it wishes to generate--it must listen to music, it must read essays, it must view art--then, using algorithms designed for its specific purpose, the program will create content based on the common patterns detected in the material it has assessed. It is important to remember, however, that generative AIs do not actually produce new content or material, but rather uses information from existing work to create a reasonable average of the whole.
Terminology
What Do I Need to Know?
What Do I Need to Know?
The following terms provide a starting point for developing AI literacy and understanding fundamental concepts in the field of artificial intelligence:
Artificial Intelligence (AI): The field of computer science that focuses on creating intelligent machines capable of performing tasks that typically require human intelligence.
Machine Learning (ML): A subset of AI that involves the development of algorithms and models that enable machines to learn from data and make predictions or decisions without being explicitly programmed.
Deep Learning: A subfield of machine learning that uses neural networks with multiple layers to learn and extract complex patterns and representations from data.
Neural Network: A computational model inspired by the structure of the human brain, composed of interconnected nodes (neurons) that process and transmit information.
Algorithm: A step-by-step procedure or set of rules followed by a computer program to solve a problem or accomplish a specific task.
Training Data: The data used to train an AI model or algorithm. It consists of input examples and corresponding desired outputs or labels.
Model: A representation of patterns or relationships learned from training data. It can be used to make predictions or classify new, unseen data.
Supervised Learning: A type of machine learning where the model is trained on labeled data, with known inputs and corresponding desired outputs.
Unsupervised Learning: A type of machine learning where the model learns from unlabeled data, seeking to identify patterns, clusters, or relationships in the data.
Reinforcement Learning: A type of machine learning where an agent learns to make decisions and take actions in an environment, receiving feedback in the form of rewards or penalties.
Natural Language Processing (NLP): The field of AI that focuses on enabling computers to understand, interpret, and generate human language.
Computer Vision: The field of AI that involves the development of algorithms and systems capable of analyzing and understanding visual data, such as images or videos.
Bias: Systematic errors or prejudices in AI models that can lead to unfair or discriminatory outcomes. Bias can arise from biased training data or algorithmic design.
Ethics in AI: The study and consideration of the moral and ethical implications of AI development and deployment, including issues such as privacy, bias, transparency, and accountability.
Explainable AI: The ability of AI systems to provide understandable explanations or justifications for their decisions or predictions, enabling humans to understand and trust the AI's reasoning.
Machine Learning
Machine Learning
A type of artificial intelligence that involves the use of algorithms and statistical models to enable computer systems to learn and improve from experience without being explicitly programmed—in other words, machine learning allows computer systems to automatically recognize patterns and relationships in data and use that information to make predictions or take actions (OpenAI, 2023).
Large Language Model
Large Language Model
The process of training artificial intelligence models on massive amounts of text data to enable them to understand and generate natural language at scale—this involves using machine learning algorithms to analyze and identify patterns in large datasets of text, such as books, articles, and websites, and then using this information to develop AI models that can understand and generate natural language (OpenAI, 2023).
Neural Network
Neural Network
An algorithm that is designed to recognize patterns in data, inspired by the structure and function of biological neurons in the brain—it is a set of interconnected nodes, or artificial neurons, that work together to perform a specific task such as image or speech recognition, natural language processing, or prediction—each neuron receives input data, processes it through a set of mathematical operations, and produces output that is passed on to the next neuron in the network (OpenAI, 2023).
Two Major Generative-AI Models
Generative Adversarial Network (GAN)
Uses two connected neural networks—one to create and one to test
Generator: generates synthetic data (e.g., image, voice, & video)
Discriminator: examines the synthetic data to determine whether it is authentic or fake
The feedback loop continues until the synthetic data is recognized as “real”
Primarily used for voice generation, graphics, and video (Baidoo-Anu & Anash, 2023, p. 3)
Examples: Google, Nvidia, Facebook, OpenAI, Uber, IBM, Microsoft
Generative Pre-trained Transformative (GPT)
Uses a large amount of publicly available digital content data to read and produce human-like text in several language and can exhibit creativity in writing from a paragraph to a full research article (or near convincing) on any topic
Models are able to engage customers in human-like conversations such as customer-service chatbots or fictional characters in video games
Primarily used for voice text generation (Baidoo-Anu & Anash, 2023, p. 3)
Examples: Google, Salesforce, Facebook, OpenAI, Amazon, IBM, Microsoft
How it Works
This graphic was created using MidJourney, an AI generator that creates images from language descriptions (MidJourney, 2023).
AI Prompt
AI Prompt
Create a technical illustration of a woman sitting behind a desktop computer on a long table, isometric view, 3D rendering, realistic 4k
Common Generative AI Tools
ChatGPT (OpenAI) – in response to a prompt can output text in any form—prose, poetry, computer code, etc. (uses GPT4)
Gemini (Google) – can enhance various applications by generating text, images, and other content
Microsoft CoPilot – integrates with Microsoft 365 applications to provide real-time assistance, enhancing user creativity and efficiency
DALL-E 2 (OpenAI) – takes text and transform it into graphics
Colossyan – AI video and audio tools
PhotoMath – has character recognition software to respond to mathematical inquiries
HyperWrite – an AI-powered writing assistant
WordTune – helps improve the clarity and effectiveness of writing
Soundraw – AI music generator that allows users to create royalty-free music based on the instruments they select, mood they prefer, and the length of the track
Murf – text-to-speech tool that allows user to create natural-sounding synthetic vocal recordings in 15 different languages with over 100 different dialects and voices
Duolingo – language-learning platform that uses AI to tailor lessons and practice exercises based on individual user progress and learning styles
Quizlet – A study tool that uses AI to create personalized study sets and quizzes
Grammarly – AI writing assistant that offers real-time grammar, punctuation, and style suggestions
Practical Applications
Grading and Assessment: automate grading and assessment tasks, such as multiple-choice quizzes, freeing up instructors to focus on more complex assignments and providing faster and more consistent feedback to students
Personalized Learning: generate personalized learning content, such as quizzes, assignments, and readings, based on each student's individual learning style and pace
Research and Analysis: analyze large data sets and generate insights and predictions that can inform research
Content Creation: create content for online courses, such as interactive simulations, videos, and animations
Administrative Tasks: automate administrative tasks, such as scheduling and course planning, allowing instructors to focus on teaching and research
References
Baidoo-Anu, D. & Owusu Ansah, L. (2023). Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in promoting teaching and
learning. Social Science Research Network (SSRN). https://ssrn.com/abstract=4337484
MidJourney, Inc. (2023). MidJourney (version 5) {text-to-image model}. https://www.midjourney.com/
OpenAI. (2023). ChatGPT (April 23 version) [Large language model]. https://chat.openai.com/chat
© 2023 PG AI Task Force | Webmaster: William Ashley Johnson | Contact: wjohnson3@purdueglobal.edu
Page updated
Report abuse