GenAI Studio: News, Tools, and Teaching & Learning FAQs
These sixty minute, weekly sessions – facilitated by Technologists and Pedagogy Experts from the CTLT – are designed for faculty and staff at UBC who are using, or thinking about using, Generative AI tools as part of their teaching, researching, or daily work. Each week we discuss the news of the week, highlight a specific tool for use within teaching and learning, and then hold a question and answer session for attendees.
They run on Zoom every Wednesday from 1pm – 2pm and you can register for upcoming events on the CTLT Events Website.
News of the Week
Each week we discuss several new items that happened in the Generative AI space over the past 7 days. There’s usually a flood of new AI-adjacent news every week – as this industry is moving so fast – so we highlight news articles which are relevant to the UBC community.
In this week’s tech news, Google released a new update for the Audio Overview feature on NotebookLM, allowing users to customize AI-generated audio recordings to focus specific topics or adjust the level of complexity. Simon Willison demonstrates an example of this new feature using pelicans. Similarly, Meta released NotebookLlama, which closely mimics the podcast generation feature of NotebookLM, but uses their own Llama models and is said to be “open”. The Open Source Initiative (OSI) released a new official definition of “open-source” for AI, which requires that the training data be released (among other things) to be considered “open-source”. Google releases its AI watermarking system called SynthID open-source, a tool that embeds watermarks into AI-generated content. Anthropic released a new feature called Computer Use, and updated the Claude 3.5 Sonnet and Haiku models, while Perplexity has released its AI-powered search App that is native to macOS. Penguin Random House has updated their copyright pages to prevent AI tools using their data and intellectual property for training. Finally, experts express concerns as OpenAI’s transcription tool has been found to generate hallucinations, but hospitals continue to use the tool.
Here’s this week’s news:
Google’s NotebookLM Update
Google has released a new update for NotebookLM, now equipped with one of its most requested features. The update improves its functionality on the Audio Overview feature, which now allows users to customize the AI-generated podcasts. Users can give directions to have the “AI hosts” focus on specific topics that is provided in the source, and to even mention certain stories and analogies in the resulting podcast.
Read more about the update, and try out NotebookLM here.
Simon Willison’s Review of NotebookLM
This blog post by Simon Willison, an AI researcher, demonstrates the use of NotebookLM’s new customization feature for Audio Overview. Taking a link to a post about video scraping, and the topic of pelicans, the demonstration shows the AI hosts integrate topic-specific analogies and anecdotes into the resulting audio recording. This demonstration shows a fun example on how NotebookLM achieves its customization on specific topics and expertise levels in its Audio Overview features.
Meta Releases an ‘open’ Version of Google’s Podcast Generator
Meta has released an open-source version of Google’s podcast generator, enabling developers to create custom, AI-driven podcast content. Named NotebookLlama, this tool creates podcast style recordings by generating a transcript from a file, transforming the transcript into a podcast-style script, and finally by feeding this script into a text-to-speech model. This tool lowers barriers for podcasters to access advanced audio technology without proprietary restrictions.
OSI’s New Definition on Open-source AI
The Open Source Initiative (OSI) has released its official definition for open-source AI, aiming to establish clearer standards amid debates over transparency and ethics. Under this new definition, model training data and its codebase must be made available to be considered open-source, which would disqualify large models, such as Meta’s AI models from being open-source. This definition outlines requirements for openness, transparency, and ethical development, crucial for AI’s evolution in open-source ecosystems.
Google Releases AI Watermarking Tech as Free Open Source Tool
Google is offering its AI watermarking technology, SynthID, as a free, open-source toolkit. This new version, designed to embed undetectable watermarks in AI-generated images, aims to curb misinformation by enabling easy identification of AI-created content across various platforms.
Anthropic Introduces Computer Use, and New Claude 3.5 Sonnet & Haiku
Anthropic has launched new updates for Claude 3.5 Sonnet and Haiku, the latest upgrade in its AI model lineup, optimized for complex language tasks and extended usage on computer systems. The update also features a new experimental capability called “computer use”, which allows Claude to directly interact with the user’s computer to accomplish tasks, including actions such as “looking at a screen, moving a cursor, clicking buttons, and typing text.”
Check out Anthropic’s full announcement here.
Perplexity Launches Native macOS App for AI-Powered Search
Perplexity AI, a conversational AI tool designed to streamline complex question-answering, is now available on macOS. The app provides Mac users with an intuitive AI interface for quick, accurate responses, optimized for macOS workflows.
Penguin Random House Underscores Copyright Protection in AI Rebuff
Penguin Random House has reiterated its commitment to copyright protection, by making changes to its copyright pages. The changes clarify that AI systems should not use the publisher’s content without permission for training purposes. This stance reinforces industry concerns around intellectual property and the ethical use of content in AI training.
Concerns over Hallucinations and Use of OpenAI’s Transcription Tool
Recent reports highlight that AI transcription tools in hospitals occasionally generate “hallucinations,” adding false information into patient records. Despite warning messages from OpenAI discouraging use of these tools in “high-risk domains”, a large number of medical workers continue to utilize tools as they receive pressure to decrease costs. This issue raises concerns about the reliability and safety of AI-driven documentation in critical healthcare settings.
Tools of the Week
Tool of the Week: AnswersAI
What is AnswersAI?
AnswersAI is a new Q&A-focused AI tool, offering personalized responses to user queries on a variety of topics. It provides rapid, precise answers and is optimized for quick access to information through both web and mobile interfaces.
How is it used?
Users can ask questions directly through the AnswersAI platform, and the tool’s algorithms process these queries to generate detailed and relevant responses. It’s designed to handle a wide range of topics, from academic inquiries to general knowledge questions, with integration options for use on social media. This tool can also listen directly to live lectures or recorded videos to creates notes or answer questions in real-time.
What is it used for?
AnswersAI is suited for individuals seeking quick, concise answers across diverse subjects, making it a valuable tool for both casual inquiries and professional research. It offers convenient solutions for anyone needing quick answers at their fingertips. Though it comes with its concerns on academic integrity, if used responsibly, this tool could increase accessibility for educational resources and allow learners to engage deeper with teaching material.
Check out a demonstration here, and try out AnswersAI.
Tool of the Week: Lecture.new
What is Lecture.new?
Lecture.new is a streamlined online tool for creating lectures, presentations, and educational content quickly. Designed with educators in mind, it aims to simplify content creation by automating layout and presentation design, allowing users to focus more on instructional material.
How is it used?
Users can begin by inputting their content or uploading existing materials, after which the tool organizes the input into a professional format with customizable elements. This includes slide generation, graphics embedding, and outline creation to facilitate smooth and engaging lectures.
What is it used for?
As the tool offers various formats for organizing and engaging with education materials, this tool is ideal for educators, trainers, and content creators. Lecture.new is used to create dynamic presentations and lectures efficiently, supporting effective teaching and engaging learning experiences.
For more information, check out Lecture.new.
Without a PIA, instructors cannot require students use the tool or service without providing alternatives that do not require use of student private information
Questions and Answers
Each studio ends with a question and answer session whereby attendees can ask questions of the pedagogy experts and technologists who facilitate the sessions. We have published a full FAQ section on this site. If you have other questions about GenAI usage, please get in touch.
-
Assessment Design using Generative AI
Generative AI is reshaping assessment design, requiring faculty to adapt assignments to maintain academic integrity. The GENAI Assessment Scale guides AI use in coursework, from study aids to full collaboration, helping educators create assessments that balance AI integration with skill development, fostering critical thinking and fairness in learning.
-
How can I use GenAI in my course?
In education, the integration of GenAI offers a multitude of applications within your courses. Presented is a detailed table categorizing various use cases, outlining the specific roles they play, their pedagogical benefits, and potential risks associated with their implementation. A Complete Breakdown of each use case and the original image can be found here. At […]