Artificial Intelligence (AI) Help Guide

• Brainstorm: Whether you're trying to identify a research topic, generate search keywords, or plan tonight's dinner, generative AI tools can help you jumpstart your work and organize your thoughts.


• Summarize: Some generative AI tools, such as Microsoft Copilot and Perplexity, can summarize text, including articles, reports and email threads.


• Build templates: Generate editable templates for emails, cover letters, resumes, checklists, social media posts, flyers, presentations and more. You can adapt these templates to meet your individual needs.


• Study: Use generative AI to create study tools, like flashcards, practice quizzes, outlines, and summaries. Preparing for finals? Try prompting an AI tool to create a personalized study schedule!


• Revise and proofread: Improve your writing with AI tools like Grammarly, which corrects spelling and grammar mistakes and offers suggestions for clarity, tone, and style. Grammarly is available to all UDC students, faculty, and staff.


• Research: Discover new research and streamline your literature review using literature search, synthesis, and visualization AI tools, such as Elicit and ResearchRabbit. 


• Translate: Translate into and from different languages using AI tools like Google Translate or DeepL.


• Illustrate: Create images from text prompts using AI image generators like DALL-E 3 or Designer, which is integrated into Microsoft Copilot.


• Explore careers: Use generative AI tools to revise your resume, understand the skills needed for different jobs, and simulate practice interviews. 


• Enhance accessibility: Generative AI tools can enhance accessibility for people with disabilities. Possible uses include creating alternative text for images, generating closed captions and transcripts for audiovisual materials, and offering task support and other cognitive assistance functions. 

• Hallucinations: Generative AI tools, especially large language models (LLMs) like ChatGPT or Microsoft Copilot, are prone to "hallucinations," or giving false or misleading information. They may present this incorrect information with confidence and ambiguous or fabricated citations. 


• Inconsistencies: The outputs of generative AI can be inconsistent. Two people can enter the same prompt into the same AI tool and receive a different response or output. Generative AI tools create new content, and that content is not always reproducible. This poses challenges when predictable, repeatable behavior is required.


• Scope of training data: AI tools can only generate content based on the data they have been trained on. If the training data is limited, biased, or flawed, the AI system's outputs will reflect those shortcomings. For example, many generative AI models are trained mostly on English content from Western countries, so the content they generate might exclude diverse global viewpoints.
  
  Training data also often has a cutoff date, meaning many AI tools may lack knowledge of recent events and new developments. The current free version of ChatGPT, for instance, was trained on data last updated in September 2021, so it will struggle to respond to prompts where the latest knowledge is critical.


• Paywalled content: Literature search AI tools, like Elicit and ResearchRabbit, may help you identify research articles, but they do not have access to the full text of articles that are behind a paywall. Use your UDC credentials to log into library-subscribed databases for full access to available articles—or request an article through Interlibrary Loan if a copy is not available through UDC Library. 


• Less creativity, novelty, and nuance: Generative AI tools remix and repurpose existing data and patterns, but they lack the ability to create truly new ideas or concepts. AI systems may also struggle with understanding nuance and context, especially in more complex scenarios.


• Ethical and legal concerns: AI raises several concerns related to ethics, bias, privacy and security, copyright, labor, and its environmental impact. Learn more on our AI and Ethics page.

Presented by Glen J. Benedict, Access Services Librarian at UDC Library, on January 31, 2025.

AI is everywhere and it seems like it can do anything. In this webinar, we share a quick refresher on the fundamentals of generative AI, followed by an exploration of its potential applications and advantages. However, AI isn’t flawless. We’ll also dive into common errors and inconsistencies in gen AI, discussing how these challenges could affect your work and offering strategies to navigate them effectively.