Home
LITS

What is AI?

Background

The current generation of artificial intelligence (AI) algorithms and tools are the result of decades of groundbreaking research across cognitive science, computer science, economics, game theory, and mathematics dating back to the 1950s.

So, what do these recent advancements in AI platforms and their widespread accessibility mean for students and educators? As AI technology rapidly evolves, so do its potential applications and the ethical questions it raises. This guide offers valuable background and resources to help you navigate this exciting, fast-changing landscape!

Key Terms

Modern AI platforms have various algorithms designed to achieve specific goals. At their core, these platforms take training data, apply machine learning algorithms to “learn” from it, and then pass this knowledge on to a model that uses it to generate outputs. Below are simple definitions of key concepts related to modern AI platforms.

  • Algorithms are a process or set of rules that a computer uses to solve problems.
  • Generative AI is a type of AI system that generates text, images, or other media in response to user prompts.
  • Generative Pre-Trained Transformer (GPT) is an LLM developed by OpenAI that uses unsupervised learning.
  • Large Language Models (LLMs) such as ChatGPT apply deep neural networks to analyze huge amounts of text and use this information to respond to user prompts.
  • Machine learning is a sub-field of AI focused on the problems of designing recursive algorithms capable of learning.
  • Natural Language Processing is a branch of artificial intelligence concerned with enabling computers to understand and generate text and spoken words in the same way humans can.
  • Neural Networks are an approach to machine learning that uses many simple but densely connected algorithms to solve complex problems.
  • Deep Neural Networks employ many layers of neural networks.
  • Supervised learning is a machine learning technique in which the model authors tell the machine learning algorithm how to handle the training data in order to generate the desired output.
  • Training data are the information that is digested by a machine learning algorithm.
  • Unsupervised learning is a machine learning technique in which the algorithm creates its own labels for variables within the training data.

Uses and Limitations

Ethical Questions:

  • How can you use AI tools and maintain academic integrity (including avoiding plagiarism)?
  • Are AI tools impacting your education? Do they undermine your opportunity to learn, or do they enhance your understanding?
  • Are AI tools providing biased responses?

Generative AI can be helpful for:

  • Developing ideas for a topic and identifying related concepts.
  • Suggesting which library databases might help find sources and which keywords or search strategies to try when searching those other databases. (You can also get human help with these tasks!)
  • Providing suggestions for improving writing. (You can also get human help with writing from the college’s Writing Center!)
  • Creating translations and asking questions in another language.
  • Assisting with computer coding tasks. (Always understand and test prior to use)

Limitations of current generative AI:

  • Reliability: They can’t assess the accuracy or logical consistency of the text they generate. They can make up facts and credible-looking sources, complete with citations to nonexistent books and articles. This phenomenon is often referred to as a “hallucination” or “confabulation.”
  • Bias: They perpetuate systemic biases. They are trained on datasets scraped from the Internet, which includes all the racist, sexist, ableist, and otherwise discriminatory language and images found there.
  • Legal Violations: They can infringe on copyright and intellectual property.

What can you do? 

  • Develop your fact-checking skills to spot and correct errors that generative AI tools might make.
  • Verify citations and sources provided by generative AI tools.
  • Develop depth of knowledge in your discipline or areas of interest to identify erroneous or biased information.
  • Ask other human experts like your professors or teaching assistants.
  • In general, learn more about the tech tools you use. Find out how they are trained, their purpose, and their limitations.

Ethical Use

Familiarize yourself with Trinity’s AI Principles. If you use a generative AI tool for writing, be transparent. Acknowledge your uses of the tool (such as editing your writing or translating words) within your paper, in a note, or another suitable location like an appendix.

Ethical Question:

  • Should generative AIs be credited the same way as human authors? Citations are used to credit other people’s work and give readers a path to the sources used. A reader can then find and look at those sources to make their judgment on things like authority and accuracy. AI tools generate text in human language but may not accurately identify specific sources used. However, they can fabricate convincing citations to sources that don’t exist.

What can you do?

  • Cite a generative AI tool when you paraphrase, quote, or incorporate other content (e.g., text, images, data) from it into your work.
  • Providing a record of chat prompts and responses can also be helpful for someone assessing your work. Take screenshots or provide share links (now available for ChatGPT and Gemini).
  • When it comes to finding reliable sources, use search tools and resources featured in our library guides. They are still the most effective way to find and connect to authentic published sources.

Using Copyrighted Materials

Generative AI tools are trained on collections of material gathered from many places. Some AI image and text generation tools have been trained on material scraped from web pages without the consent or knowledge of the web page owners. User-entered information or uploading of PDFs into chat prompts might also contribute to an AI tool’s training and reuse in its responses to other user prompts.

Several ongoing lawsuits claim that using artists’ or writers’ content without permission to train generative AI infringes copyright. The courts will have to decide whether this is justified fair use or infringement.

Ethical Question:

  • Do generative AI tools violate copyright? Copyright is a form of intellectual property protection provided by the laws of the United States to the [human] creators of “original works of authorship.”

What can you do? When using AI tools, be cautious about entering or uploading copyrighted material into a prompt. Here are a few scenarios to consider:

  • Read the Terms of Service or User Agreements to learn how your information is used and any privacy options.
  • If you enter or upload copyrighted material into a generative AI tool, you may be protected by Fair Use if you are using the information for educational purposes and in a way that follows the College’s Code of Academic Integrity. However, Fair Use can be superseded by the library’s license agreements with providers who restrict how content, such as journal articles, can be used or shared with an AI tool.
  • Don’t enter or upload copyrighted material into the prompts of any generative AI tool if you are not doing so for educational purposes. This may violate copyright.
  • Be cautious about entering your work into a generative AI prompt, as that material may be shared with others without acknowledging your authorship.

Protecting Data Privacy

Information shared with AI tools using default settings is often not private and could expose proprietary or sensitive information to unauthorized parties. As with any technology, data breaches are also a risk.

What can you do?

  • Read the Terms of Service to find out how your information is used and if there are options to make it private.
  • Don’t share personally identifiable information such as addresses, emails, birthdays, passwords, etc. Also, while chat AIs can seem conversational, you do not want to divulge any private information you might share with a therapist, doctor, or other trusted human confidant.
  • Don’t enter workplace (college) data classified as level 2, 3, or 4, including non-public research data, financial data, human resources records, student records, etc.) into publicly available generative AI tools in accordance with the College’s Data Classification Standard and Information Security Policy.
  • If using AI tools to summarize meeting minutes or write emails, remove any information that could identify you or your colleagues.
  • Learn more about Digital Privacy.