Image generated by Craiyon from the prompt "University of Iowa"

Critical thinking is especially important when evaluating and interpreting AI-generated results. If we look at these AI-generated images, and compare them to what the University of Iowa actually looks like, there is no question that these results are not at all accurate, including the logo, building designs and icons.

There is no documentation of how and why this AI-generated tool built these pictures or the accuracy in sourcing. This can make it challenging further to understand what is right and wrong while using AI. The key is to check your sources, use multiple sources for research and back-up your content through references. As we navigate the changing environment of AI, it is important to note that we are all learning, re-learning and working through the accuracy and effectiveness of different generative artificial intelligence tools. 

Reference: University of San Diego Libraries

Using generative AI tools effectively requires the user to know the right questions to ask, and how to phrase them for the best results. Vague or generic questions generate vague or generic results. (In other words, garbage in, garbage out.) It's important to remember to use generative AI tools wisely and that feeding personal information or works under copyright into an AI system should be avoided.

Tips for crafting prompts to get the best results from chatbots:

• Write clear instructions: be specific about the format of the output - number of words or paragraphs, writing style or tone, reading level, formatting such as bullet points, a table, html, css, etc. Treat the chatbot like a brand new assistant who is eager to please but doesn't know what they don't know.
• Instruct the chatbot to take on an expert role: for example, "act as a mathematician" or "take on the role of a professional news blogger" or "I want you to act as a Linux terminal."
• Provide the chatbot with examples or a reference text.
• Split complex tasks into simpler subtasks. 
• Beware hallucinations, or the chatbot confidently stating incorrect or made up information. If you are unfamiliar with the topic, check the chatbot's work. ChatGPT and other chatbots are notorious for making up citations to sources that simply do not exist. CoPilot, LABS.Google (Google Search Generative Experience), or Perplexity may be better choices because they provide links to the websites they claim as information sources.  

• Learn Prompting online course/ebook

  The definitive guide to prompt engineering. Extensive but easy-to-read online guide to crafting prompts. Each chapter is only a few paragraphs and includes illustrations and hands-on examples. Great as a reference - does not need to be read all at once or sequentially.

• AI-Based Literature Review Tools: Approaches to Write or Fine Tune Prompts (Texas A&M University Libraries guide)

  Thorough overview with examples of different types of prompts. Serves as a good one-page cheatsheet to Learn Prompting.

• GPT best practices (OpenAI)

  OpenAI's guide to strategies and tactics for getting better results from GPTs offers six strategies for getting better results

• Quick Guide: AI Prompt Engineering infographic (by Jeri Hurd, Branksome Hall Asia)

• Get the Best From ChatGPT With These Golden Prompts (The New York Times: On Tech: A.I.)

• Generative AI Is Only as Good as the Prompt You Give It (Inside Higher Ed, 4/26/2023)

• Prompt Engineering for ChatGPT (Coursera course)

• Awesome ChatGPT Prompts (GitHub repository)

  Collection of pre-written prompts.

Many tech reviewers have published comparisons between ChatGPT and CoPilot, reviewing the responses of each to a variety of prompts. The paid subscription version of ChatGPT, powered by GPT-4, almost always scores the highest. Bing (now referred to as CoPilot in Edge), which is also powered by GPT-4 plus Bing web search, often scores well and has the added bonuses of being free and linking back to websites containing the information it provides. 

• Chatbot Arena

  Created by researchers at UC San Diego and UC Berkeley, Chatbot Arena is a benchmark platform for large language models (LLMs) that features anonymous, randomized battles in a crowdsourced manner.

• The best AI chatbots: ChatGPT and other noteworthy alternatives (ZDNet, 6/21/2023)

  Compares Bing (selected as best overall), ChatGPT, Perplexity AI, Jasper, YouChat, Chatsonic by Writesonic, Google Bard, and Socratic by Google.

• Awesome-LLM

  GitHub repo of featuring curated list of papers about large language models, frameworks for LLM training, tools to deploy LLM, courses and tutorials about LLM and all publicly available LLM checkpoints and APIs:

• Open LLM Leaderboard

  The Open LLM Leaderboard aims to track, rank and evaluate LLMs and chatbots as they are released. Anyone from the community can submit a model for automated evaluation on the GPU cluster.

• ChatGPT vs. Microsoft Copilot vs. Gemini: Which is the best AI chatbot? (ZDNet, 2/9/2024)

• Conversational AIs for Business Reference: How They Work and What They Do Comparison

• Perplexity

Perplexity AI is an AI-powered search engine and chatbot that leverages advanced technologies like natural language processing (NLP) and machine learning to provide accurate and comprehensive answers to user queries. Designed to search the web in real-time, it offers up-to-date information on a wide range of topics. With an intuitive user interface, Perplexity AI is a valuable resource for students, researchers, and anyone seeking precise answers supported by citations.

• CoPilot

  CoPilot is only available in Microsoft Edge and stand-alone mobile app. Chatbot powered by Microsoft's proprietary Prometheus, which combines GPT-4 (which also powers the subscription version of ChatGPT) and Bing web search.

• LABS.Google or Google SGE (Search Generative Experience)

  Google Search Generative Experience (SGE) is an AI-driven feature that generates interactive AI snippets to provide users with accurate and relevant answers to their queries, along with links to corroborating sources AI generated summary. Also provides the standard Google search results links below the AI summary. Join the waitlist in Google Labs.

• Consensus

  Consensus is a search engine that uses AI to extract and distill findings from peer-reviewed sources. Subject matter coverage ranges from medical research and physics to social sciences and economics. Utilizes the Semantic Scholar dataset.

• Elicit

  Helps automate research workflows, like creating literature reviews, brainstorming, summarization, and text classification. Utilizes the Semantic Scholar dataset.

• Research Rabbit

  Provide a seed article that allows you to retrieve recommended papers, visualize networks of papers and authors, and get alerts about additional relevant research. It integrates with Zotero and allows for collaborative research sharing.

• Semantic Scholar

  A project at the Allen Institute for AI, it indexes over 200 million academic papers sourced from publisher partnerships, data providers, and web crawls.

• JSTOR Text Analyzer

  Built by JSTOR Labs, you can search for content on JSTOR by uploading a document. Supports English, Arabic, (simplified) Chinese, Dutch, French, German, Hebrew, Italian, Japanese, Korean, Polish, Portuguese, Russian, Spanish, and Turkish.

  • 

• Litmaps

Litmaps uses the citation network of discoverable literature to build large-scale searching. You can use Litmaps to get alerted when new papers of current topic focus, discover key authors in field, find research gaps and share with other researchers.

• ChatGPT and Bing Chat Generative AI Legal Research Guide (University of Arizona Law Library)