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.

Here’s this week’s news:

1. Google Genie, developed by Google DeepMind, is an innovative AI platform capable of generating interactive 2D video games from text or image prompts. It operates using a ‘world model’ trained on 200,000 hours of video footage from 2D platformer games, allowing it to generate video games based on minimal input. While Genie’s visual quality and access are currently limited, its potential in game development and entertainment is significant.
   
2. In a British Columbia court case, Vancouver lawyer Chong Ke used AI, specifically ChatGPT, to prepare legal material which included citations to nonexistent cases, a situation described as an AI “hallucination.” Ke acknowledged the mistake and its implications, stressing the need for awareness of the risks associated with using AI in legal work. This incident has prompted discussions among legal experts and law societies about the cautious and skeptical use of AI in law, highlighting the importance of human expertise and professionalism in the justice system. The case serves as a wake-up call about the limitations and necessary critical approach towards AI in high-stakes fields like law.
   
3. Anthropic has announced the Claude 3 model family, comprising three AI models: Claude 3 Haiku, Claude 3 Sonnet, and Claude 3 Opus. These models, offering varied capabilities, set new benchmarks in AI performance across multiple cognitive tasks. They feature improved speed, accuracy, fewer refusals, enhanced vision capabilities, and long-context processing. Opus and Sonnet are currently available, with Haiku set to launch soon.
   
4. Google has collaborated with Stack Overflow to integrate Stack Overflow’s knowledge base into Gemini for Google Cloud. This partnership allows for validated Stack Overflow answers to be accessed directly in the Google Cloud console, enhancing the developer experience by providing reliable, expert-generated content alongside Google-specific responses. This integration is a significant step towards blending AI and human expertise in developer workflows.
   
5. A study published in Scientific Reports demonstrates that GPT-4 outperformed humans in tests of divergent thinking, a key indicator of creative potential. The study involved tasks such as coming up with creative uses for everyday objects and imagining outcomes of hypothetical scenarios. GPT-4 showed more originality and elaboration in its responses compared to human participants, suggesting a higher creative potential in specific AI models.
   
6. The Curious Decline of Linguistic Diversity: Training Language Models on Synthetic Text. This research paper explores the consequences of training large language models (LLMs) on synthetic data generated by their predecessors. It reveals a marked decrease in linguistic diversity across various natural language generation tasks as models are recursively trained on such data. This trend underscores potential risks associated with this training methodology, particularly regarding the preservation of linguistic richness in language models.
   
7. SunoAI, a company based in Cambridge, MA, is paving the way for a future where anyone can create great music, regardless of their musical background. With their latest offering, V3 Alpha Access, SunoAI invites users, from casual enthusiasts to professional artists, to test and improve this innovative music generation tool.
   
8. OpenAI has successfully disrupted the activities of five state-affiliated threat actors who attempted to misuse AI services. These actors, linked to China, Iran, North Korea, and Russia, used AI services for tasks like open-source research, scripting, and content creation for phishing campaigns. OpenAI, in partnership with Microsoft Threat Intelligence, identified and terminated the associated accounts, showcasing their commitment to combatting the misuse of AI technologies.
   
9. Semantic Scholar is a free, AI-powered research tool for scientific literature, based at the Allen Institute for AI. It provides a comprehensive database for searching scientific papers across various fields. The platform features tools like Semantic Reader, offering an augmented reading experience that makes scientific literature more accessible and contextually enriched. 
   
10. Humata AI offers an AI-powered tool designed for working with data files, particularly PDFs. It enables users to upload documents, which are then processed by the AI to provide summaries, comparisons, and answers to queries based on the content. Humata AI features include unlimited file uploads, citation highlights, unlimited questions, and embeddable web integration. It serves as a tool to efficiently manage and interact with large volumes of documents, enhancing research and information retrieval

Tool of the Week

Each week we demonstrate a Generative AI tool that can be used within teaching and learning. The GenAI space is evolving rapidly, and as such we demo new tools or new ways people use those tools.

As a reminder not all tools we showcase have successfully been through the PIA process at UBC.

This week we actually have 2 Tools of the Week: Semantic Scholar And Humata AI

Semantic Scholar is an AI-powered research tool that streamlines the discovery of scholarly articles. It uses advanced algorithms to analyze, categorize, and summarize academic papers, making it easier for users to find relevant research and understand key trends in their field. This tool is particularly effective in helping users navigate the vast ocean of academic literature, providing insights into the most critical aspects of research papers, such as methodologies, results, and conclusions.

Humata AI, complementing Semantic Scholar, offers an interactive query-answering feature. Once relevant articles are identified through Semantic Scholar, users can utilize Humata AI to ask specific, targeted questions about the content of these articles. Humata AI processes these queries using its advanced language model, delivering concise, accurate answers drawn from the article’s content, including citing specific sections or data for reference.

The combination of Semantic Scholar’s efficient literature discovery and Humata AI’s precise query-answering capabilities offers a synergistic approach to academic research. This integration allows users to not only identify the most pertinent articles quickly but also to extract and understand complex information from these articles in a more interactive and user-friendly manner.

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.