Bully's Basics

Artificial Intelligence (AI) is the ability of a computer to process and learn in ways that are similar to a human being. AI can learn from data and experiences. AI computers can also do things that normally require human intelligence, like recognizing images, playing games, or even driving cars. AI computers can perform some tasks faster than a human can, such as analyzing large amounts of data. However comforting or disconcerting, there are certain limitations to AI. The human brain has the ability to think, reason, and understand. AI cannot match these abilities, and may never be able. Humans also have emotions, creativity, and a level of consciousness that Artificial Intelligence just cannot reproduce at this time. Think of AI as a really smart calculator or a problem-solving tool that handles certain tasks really well. AI just does not have the overall intelligence and understanding that humans do though.

DISCLAIMER: It is up to each student on campus to conform to the AI policies stated by their instructors in class. This guide does not condone the misuse of AI in any way. Our hope is to educate users on the basics of Artificial Intelligence so that an understanding of ethical uses can be a campus-side conversation.

AI computers can learn by being trained on what are called Large Language Models (LLM’s). LLM’s are not just words; they can be images, movies, or even audio files. If it is a set of data, then the AI system can learn from it or train with it.

The AI computer is given a set of rules and instructions, just like a human would need if doing something for the first time. Data is inputted into the the AI computer, and the computer uses the rules to determine what it learns. Algorithms are programmed into the AI computer; this allows the computer to analyze and look for patterns and relationships in the data. The AI computer then might adjust the rules and instructions based on what it has learned from the data. The more data the AI computer is given, the more it can learn. AI Systems are doing this through a process called Machine Learning (ML).

Real-World Example
Imagine that you want to teach a computer to identify images of cats or dogs. You would feed images of cats into the computer and tell the computer that the images are cats. Then you would feed images of dogs into the computer and the computer that the images are dogs. This is the Large Language Model (LLM). Then you would feed it images of different cats and dogs than the images the computer was trained on. The computer would use what it learned about cats and dogs already to “guess” if the new images were cats or dogs. The more the AI computer “guesses” correctly, the more it learns. This is Machine Learning (ML).

This is an example of how an AI system would use this image to "learn" what a building is, what a human is, what a bicycle is, and what a bag is.

You are already using AI in so many ways! In fact, you have probably been using it for years, or even your entire life (depending on your age). There are so many things that we do every day that require Artificial Intelligence. Here are just some examples of AI in everyday life:

Siri, Alexa, Hey Google (Personal Assistants)
Natural Language Processing (NLP) is what allows AI systems to hear you and respond to you.

Social Media
Social Media platforms use AI to examine your behavior on their platform and then suggest ads, posts, and other personalized experiences.

Travel and Navigation
Apps like Google Maps use AI to monitor traffic and weather and suggest ways to avoid traffic congestion or accidents.

Healthcare
AI is used in medical diagnosis, medical imaging, and medication research.

Banking
If you deposit a check with your mobile app, AI scans the image to read the handwriting (or typing) and verifies the amount.

Smart Home Devices
Smart devices such as thermostats, light bulbs, or video doorbells use AI to learn the preferences of the user and adjust settings.

Content Generation
Apps like chatGPT, Bard, DALL-E, and Adobe Firefly all use content generation. You give the app a description of something, usually in text, that you want it to create, and it does. Items created can include a variety of text, computer code, digital images, or with some software, even voices.

Apps like chatGPT can do a number of these generative AI tasks. Since chatGPT was written as open-source software, meaning anyone can use, copy, or modify it, we have seen an eruption of these AI Assistants in 2023. You can now use AI to plan a vacation, look for lasagna recipes online, or even motivate and restructure your running routine. Generative AI is not only captured imaginations on a global scale, but it has also raised questions of ethics.

The most important thing to remember about artificial intelligence is that it is not perfect. It has many limitations.

Here are some key limitations that were put forth by chatGPT.

Lack of Common Sense: AI systems often struggle with common-sense reasoning. They may excel at specific tasks but lack a deep understanding of context and human intuition.

Data Dependency: Most AI models require large amounts of high-quality data to learn and perform well. They may not function effectively if the data is biased, incomplete, or unrepresentative.

Interpretability: AI models like deep neural networks are often seen as "black boxes." It can be challenging to understand how they make decisions, which can be a problem in critical applications like healthcare or law.

Narrow Expertise: AI systems are typically designed for specific tasks. They lack the versatility and adaptability of the human mind, which can seamlessly switch between various cognitive tasks.

Ethical Concerns: AI can perpetuate biases present in training data, leading to unfair or discriminatory outcomes. Ensuring AI systems are ethical and unbiased is a significant challenge.

Resource Intensive: Developing and training AI models can require significant computational power and energy resources, making them inaccessible to smaller organizations.

No Genuine Understanding: AI systems do not possess true comprehension or consciousness. They process data and generate responses based on patterns but lack genuine understanding or consciousness.

Limited Creativity: While AI can generate content like art or music, it often replicates existing styles rather than creating entirely novel works.

Vulnerability to Adversarial Attacks: AI models can be susceptible to manipulation through carefully crafted input, known as adversarial attacks, which can have security implications.

High Costs: Developing and maintaining AI systems can be expensive, limiting their adoption in some industries or regions.

Dependence on Human Oversight: AI systems require human supervision and intervention to correct errors and ensure they align with human values and ethics.

These limitations highlight the ongoing challenges in AI research and development. While AI has made significant progress, it's essential to be aware of these constraints when considering its applications and potential impact.

Artificial Intelligence is a powerful tool that can do all sorts of things using computer systems. But it's important to think about the ethical side of how we use AI in our society. This is subject of much conversation and debate. How do you write policies, rules, and guidelines for a technology that changes almost daily? In an educational setting, how do educators ensure that students, staff, and faculty are using AI in ways that are ethical? Who gets to decide these things?

Here are some ethical points to consider in conversations about the ethics of Artificial Intelligence.

• Fairness and Social Impact: AI can influence our society in significant ways, from job opportunities to access to education and healthcare. An ethical concern is making sure AI is used fairly and doesn't create or reinforce inequalities. What guidelines or policies need to be put in place to ensure the integrity of an AI system? Who is responsible for ensuring that ethical methods and measures are followed?
• Plagiarism and AI: Is it plagiarism to use AI to generate homework, movies, books, art, or any other type of creative work? Does plagiarism also count towards non-creative work if AI is used? At what point is the AI system doing the work that a human is supposed to do?
• Ownership of Self: Recently, AI has used voice generation to "imitate" humans who are either dead or alive. The Beatles have recently released a "new" song with John Lennon on vocals, thanks to AI technology. Who has the right to determine whose voice or likeness can be used? How do we stop AI systems from using our likeness, our voice, or our creative content? Who speaks for the dead in these considerations? 
• Privacy and Surveillance: AI often deals with a lot of data about people. Think about how social media platforms use AI to show you personalized content. How much data can be collected? Is permission needed to collect and distribute AI data? Is it really harmful for an AI system to even analyze or store data? Who has access to it? How do we ensure that the AI system will respect our privacy?
• Transparency and Accountability: AI systems can make important decisions, like loan approvals or even legal judgments. How do we make these AI systems transparent? How do we govern the "decision-making" process of these decisions? Do we need to govern them? What accountability is there for mistakes or AI hallucinations?
• Creative Expression: Creativity and expression are highly valued. Ethical concerns related to AI include issues like plagiarism when AI generates content, and whether AI can genuinely create art or literature. Is it ethical for an AI system to take any element from an image or painting to use in an AI generative creation? Who then owns the copyright over the material? How does this affect existing policies in regards to copyright and trademarked content?
• Human Identity: AI raises questions about what it means to be human. Is it possible for AI to have consciousness or emotions? If AI has consciousness or emotions, does it then have rights as humans do? What makes it "alive?" How does interacting with AI affect our own humanity? Will we eventually merge with AI systems to extend our life span or intelligence beyond normal human limitations?
• Cultural and Global Context: Different cultures have different values and perspectives. Ethical considerations around AI may vary around the world, and it's important to respect diverse viewpoints. Are AI systems subject to one country's policies or every country's policies? Do we need a global force to regulate AI? Who gets to regulate it? Scientists, politicians, or someone else?

Here is a sampling of useful AI apps and resources:

Dropbox Dash - storage solutions

Tableau - Data Visualization Software Platform

TensorFlow - build machine learning projects and neural networks

Time-Hero - time management platform

chatGPT - generative AI chatbot

Bing Chat - generative AI chatbot

Google Bard - generative AI chatbot

Llama 2 - generative AI chatbot

HuggingChat - generative AI chatbot

Otter.ai - transcription service

ChatPDF - chat with any PDF

Adobe Firefly - generative AI images

DALL-E 3 - generative AI images

Midjourney - generative AI images

Bing Image Creator - generative AI images

Grammarly - generative AI writing assistance

Decktopus - presentation generator

Descript - make video and podcasts
 

The images in this Artificial Intelligence Guide were downloaded from Adobe Stock, using an educational license.