Generative AI for Teaching and Learning

Blackboard Ultra: What's New?

Generative AI offers exciting possibilities to transform teaching and learning but raises challenges around academic integrity, privacy, and bias. Here, you'll find resources on what generative AI is, how it works, and how to write effective prompts to unleash the power of AI tools. You can also find insights on how to leverage AI tools to create personalized learning experiences, enhance student engagement, and support faculty development. Explore practical applications, discover best practices, find training information, and join us in shaping the future of teaching and learning with generative AI at WPU.

What's New?

Introducing the WPU Blackboard Support Assistant! It is a tailored version of ChatGPT with extra knowledge about WPU’s policy and practices of Blackboard usage and local help files. Leveraging generative AI technology, this custom GPT offers technical support tailored specifically for the needs of WPU faculty and student Blackboard users. Click the link below to experience instant assistance at your fingertips! Please note that some of the live links provided by this GPT is intentionally disabled by OpenAI for security reasons.

Blackboard Suppport Assistant Blackboard Support Assistant

  • Generative AI Basics

    There are many definitions available. Here is one that is generated by ChatGPT 4-o:

    Generative AI refers to artificial intelligence technologies that create new content, such as text, images, audio, and video, by learning patterns from existing data. These systems utilize advanced machine learning models, including deep learning techniques like neural networks, to generate outputs that are often indistinguishable from those created by humans. Generative AI can be applied in various fields.

    The key aspects and concepts of Generative AI include:
    • Creating new content: This differentiates it from traditional AI that primarily makes predictions or classifications based on existing data.
    • Learning from existing data: Generative AI models are trained on massive datasets to understand patterns and relationships, which they then use to create new, similar content.
    • Prompt-based generation: Many generative AI systems work by responding to prompts or queries from users to guide the content creation process.
    • Underlying technologies: Deep learning techniques like advanced neural networks are crucial for the complex tasks involved.
    • Human-like outputs: The quality of generated content has reached a point where it often seems indistinguishable from human-made content.
    • Broad applications: The potential of generative AI extends across a diverse range of fields, showcasing its versatility.

      • Introduction to AI for Teachers and Students by Ethan Mollick and Lilach Mollick of Wharton School. This video is Part 1 of a five-part Practical AI for Instructors and Students course. The course explores how the Large Language models work, how to use generative AI to make teaching easier and more effective, and how students can use GenAI to improve their learning. Part 1 provides a short overview of how Large Language models (LLMs) work.

    Large Language Models (LLMs) like OpenAI’s GPT-4 are designed to be user-friendly and accessible to people without technical expertise. While it’s not necessary to understand the technical details of using LLMs, having some basic knowledge brings several benefits:

    • Help you communicate more effectively with them, allowing you to write better prompts that yield better results.
    • Help you better leverage their abilities to assist you in various tasks, from data analysis to creative writing.
    • Enable you to evaluate the information provided critically, discerning between well-supported facts and less reliable outputs.

    The core of Generative AI is the Large Language Model, and the Transformer is the architecture that enables the efficient training and functioning of these models. Below are a few excellent introductions to LLM and the Transformer for people without technical expertise:

    • A Jargon-Free Explanation of How AI Large Language Models Work by Timothy B. Lee and Sean Trott - This article is one of the best non-technical explanations of how Large Language Models (LLMs) and AI-powered chatbots such as ChatGPT work. It makes complex concepts more accessible and sheds light on the fascinating field of machine learning and language processing.
    • What Is ChatGPT Doing … and Why Does It Work? By Stephen Wolfram - Stephen Wolfram is a renowned scientist. This article explores how ChatGPT functions and why it can generate human-like text. Wolfram’s clear and engaging explanation covers neural nets, data training, probabilistic approach, word generation process, temperature parameter, and some engineering details.
    • But what is a GPT? Visual intro to transformers by Grant Sanderson (3Blue1Brown) - This video is a little more technical than Timothy Lee’s article. However, using beautiful visual animation, it explains GPT and Transformers in a remarkably intuitive and engaging way.



  • Genrative AI Tools for Faculty

    Licensed AI Tools at WPUNJ

    Zoom AI Companion is now available to all Faculty and Staff across campus. The two new features that have been approved and made available can provide hosts the ability to generate Meeting Summaries and Smart Recordings.

    • Meeting Summaries: With this feature enabled, hosts will be able to generate a summary and action items that were discussed in their meetings. The summary will be sent only to the host by way of an email after the meeting has concluded. Important Note: The AI Companion summary must be fully vetted and edited (if necessary), before being shared with others.
    • Smart Recording: Once enabled, your meeting will be recorded and saved to the cloud. You can then review your cloud recordings faster through highlights, smart chapters, summaries, next steps, and get analytics on key meeting and conversation factors.

    An AI companion, or chatbot, Microsoft Copilot (formerly Bing Chat) can help boost productivity, spark creativity, and change the way you get things done. It is available at no cost to all members of the William Paterson community with a NetID.
    You can access Copilot by logging in with your NetID on the web at copilot.microsoft.com and in Microsoft’s apps for Windows, macOS and mobile devices. You can use it to:

    • Generate or reformat text, including code
    • Answer questions using web results from Bing search
    • Create and edit images (15 boosts per day)
    • Access OpenAI’s GPT-4 and GPT-4 Turbo models during non-peak hours, including free plug-ins and GPTs
    • Access NoteBook, which helps you craft, improve and revise prompts without chat dialogue

    Copilot grants you access to Microsoft’s standard generative AI license. The university license does not grant you access to Microsoft’s premium services: Copilot Pro, Copilot for Microsoft 365 and Copilot Studio.



    Free AI Tools

    ChatGPT logo ChatGPT Claude AI logo Claude Bard logo Gemini CoPilot logo CoPilot

    Dalle2 logo Dall-e-2 CoPilot logo CoPilo Designer Freepik ai logo Freepik Leonardo ai logo Leonardo Bria AI logo Bria Openart ai logo Openart Pebblely ai logo Pebblely Playground ai logo Playground Lasco ai logo Lasco Stable Diffusion logo Stable Diffusion

    Supercreator ai logo Supercreator Runway ai video logo Runway Lumiere 3D logo Lumiere 3D shuffll logo Shuffll Fliki ai logo Fliki Synthesia ai logo Synthesia Video Reemix logo Reemix

    Altered logo Altered Voicemod logo Voicemod Krisp ai logo Krisp

    Autoregex logo Autoregex Lightning ai logo Lightning AI Github CoPilot logo Consultations

    Notion ai logo Notion AI Albus logo Albus Slides AI logo SlidesAI

    GPT-2 Output Detector logo GPT-2 Detector GLTR logo GLTR GPTZero logo GPTZero Writefull logo Writefull Content at Scale logo Content at Scale

    Paperpal logo Paperpal Scholarcy logo Scholarcy Scispace logo Scispace

    Copilot AI logo CoPilot Lesson Plan logo Lesson Plans decktopus logo Decktopus

  • Prompt Engeering

    Prompt engineering involves crafting specific instructions or questions to guide a computer program in generating desired content effectively. Within the world of generative artificial intelligence (AI), prompt engineering allows users to leverage AI models -often referred to as language models (LMs) and the natural language processing (NLP, or the instructions that allow computers to interact with humans) capabilities of those models to generate customized content that is designed according to the specifics of the prompt.

    A prompt can contain any of the following elements:

    • Task: A summary of what you want the prompt to do. Usually starts with an action verb such as, Create, Summaraize, Explore, Analyze, Write, Recommend, Argue, Explain, etc.
    • Format: The type or format of the response/output, such as Opinion, Outline, Blogpost, Spreadsheet, Press release, Leagal brief, etc.
    • Voice: Setting a specific role for a given prompt increases the likelihood of more accurate information, including like a HR director, using martketing language, in style of previous press release, respond as you are king from 17th, etc.
    • Context: Additional information or context that guides the model towards producing better outputs/responses, something like, serious, funnny, for UG students, etc.

    • The Perfect Prompt: A Prompt Engineering Cheat Sheet This comprehensive cheat sheet, created by Maximilian Vogel and published in The Generator, serves as a quick-reference guide for various aspects of prompt engineering. It is designed to be a detailed resource for both beginners and seasoned users.
    • OpenAI’s Prompt Engineering Guide This guide provides comprehensive documentation on prompt engineering, covering various techniques and best practices for crafting effective prompts. It is an excellent starting point for anyone looking to understand the fundamentals of prompt engineering.
    • GitHub’s General Guidelines for AI Prompts The repository "awesome-gpt-prompt-engineering" on GitHub is a curated list of resources and examples for prompt engineering, providing a community-driven approach to learning.
    • Top Tools for Prompt Engineering 2024: Unlock Creativity! Introduces top prompt engineering tools represent various applications, from marketplaces and models to development frameworks and optimization platforms.
    • Prompt Library from Anthropic Anthropic's Prompt Library is a valuable resource for obtaining task-specific, optimized prompts. It is particularly useful for those looking to apply prompt engineering techniques to specific use cases.

  • Genreative AI in Teaching

    Challenges

    The immediate challenge facing faculty is the fear of cheating on student assignments, tests, and exams. In fact, the ChatGPT-generated answers to assignments, tests, and exams were already used to cheat by some WPU students in the fall of 2022. This undermines academic integrity and makes it difficult for faculty to assess student understanding and progress. In the Spring 2022 semester and beyond, we can expect much more such instances. WPU needs to act and adjust to protect academic integrity quickly.

    Banning AI tools should not be an option, as it would not be effective, and it could infringe on academic freedom. Nevertheless, we can do many things. Here are a few examples:

    1. WPU should establish a task force to look at this issue and have university-wide discussions on how we should address the concerns created by these AI tools and how to adapt to the new technology.
    2. We might need new guidance on the appropriate use of AI assistance for course assignments.
    3. AI literacy. Both faculty and students need to understand the capabilities and limitations of AI tools, as well as the potential consequences of using them. Help faculty familiarize themselves with popular AI tools such as ChatGPT (and language models) to explore how to use these tools to enhance teaching methods as well as the implications of these tools.
    4. Adopt AI tools, harness their potential, and redesign teaching and learning.
    5. Take preventive measures in the current academic semester to minimize academic integrity concerns and avert the risks of the new AI tools.

    • Reiterate the honor code and academic integrity and invite students into the conversation. Openly discuss the ethical implications of using ChatGPT and other AI apps with them.
    • Develop and deploy AI writing protocols ethically and strategically within our curricula.
    • Update the syllabus. Update the statements about academic integrity and plagiarism to include ChatGPT and other AI systems.
    • Establish clear rules and expectations about the use of AI tools. Faculty should be clear in their syllabi about their policies for using ChatGPT. For example, do not allow students to use AI tools when taking quizzes, tests, and exams, and require students to appropriately acknowledge (e.g. cite AI outputs and treat them as content developed by a third party, offer rules of citation) when they do use ChatGPT in other course assignments.
    • Using proctored exam tools such as Respondus LockDown Browser and Monitor.
    • Crafting questions and designs that cannot be plagiarized by chatbots or chatbots would have difficulties to handle, such as students’ own lives and current events.
    • Encourage students’ personal reflection and connect with students’ own experiences in their assignments and discussions.
    • Giving less take-home, open-book assignments, which now seem vulnerable to chatbots. Opt for in-class assignments, handwritten papers, group work, and oral exams.
    • Redesign courses, making changes such as more oral exams, group work, and handwritten assessment.
    • Focusing on process rather than product. Scaffolding learning and allowing students to explain their thinking and make learning visible along the way are strategies that may help you confirm student originality.
    • Use AI generated text detection software, as detailed below.
    >

    Currently, several detection tools are available and can be used by faculty to identify AI-generated texts.

    • GPT-2 Output Detector
      The online demo of the GPT-2 output detector model lets you paste texts into a box and immediately see the likelihood that the text was written by AI. According to research from OpenAI, the tool has a relatively high detection rate, but “needs to be paired with metadata-based approaches, human judgment, and public education to be more effective.”
    • GLTR (giant language model test room)
      The GLTR demo enables forensic inspection of the visual footprint of a language model on input text to detect whether a text could be real or fake. It is a collaborative effort by the MIT-IBM Watson AI lab and Harvard NLP.
    • GPTZero
      Developed by Edward Tian, a Princeton University student, GPTZero measures randomness in sentences (“perplexity”) plus overall randomness (“burstiness”) to calculate the probability that the text was written by ChatGPT.
    • Watermarking (coming soon)
      The ChatGPT developer OpenAI is trying to address the problem by watermarking all ChatGPT text. “Basically, whenever GPT generates some long text, we want there to be an otherwise unnoticeable secret signal in its choices of words, which you can use to prove later that, yes, this came from GPT” (OpenAI guest researcher Scott Aaronson).
    • How about SafeAssign?
      SafeAssign checks the papers against a database of previously submitted papers, the internet, journals, and other sources to detect plagiarism. But AI-generated text is essentially unique as AI create sentences based on a statistical analysis of words in the database they are using to generate new sentences with. Hence, SafeAssign cannot detect AI-generated text unless a more sophisticated algorithm is used that is specifically designed to identify patterns and characteristics of text generated by language models.



    Opportunities

    While AI tools pose serious challenges to how we teach and how students learn, they also offer us opportunities and the potential to reimagine teaching to facilitate student learning and achieve our intended outcomes better. But this short document is focused what we need do now. I will write a more detailed report later on this part. Generally speaking, WPU could use AI tools to improve the education experience for its students and faculty in many potential ways, for example: 1) Personalized learning, 2) Grading and feedback, 3) Tutoring and support, 4) Student advisement, 5) Research, 6) Administration and management (students recruitment, registration, etc.) We should encourage faculty to conduct productive experimentation and explore pedagogical possibilities. The university might also need to establish a task force and university-wide discussion on how to take advantage of AI tools to enhance teaching and promote learning.

  • Generative AI Webinars and Workshops

    Generative AI Training Opportunities

    CTT team can assist you on adopting these feature on your teaching area. Please use one of the followig options to get connected.

    Webinars Workshops Consultations Walk-in

  • Generative AI Resources

    Recordings on the Genrative AI Sessions

    AI Assistant in Blackboard Genreative AI Tools and Prompt Engineering Generative AI and Academic Integrity