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. Stability AI announces Stable Diffusion 3, an advanced text-to-image model, currently in early preview. This model offers improvements in handling multi-subject prompts, image quality, and spelling abilities. It ranges from 800M to 8B parameters and combines a diffusion transformer architecture with flow matching. The model is in a preview phase for testing and safety evaluations before public release, with sign-ups available for early access. Stability AI emphasizes safe, responsible AI practices and is committed to making generative AI universally accessible and open source.
   
2. The paper “KOALA: Self-Attention Matters in Knowledge Distillation of Latent Diffusion Models for Memory-Efficient and Fast Image Synthesis” introduces KOALA, a distilled version of the Stable Diffusion XL model. KOALA leverages self-attention-based knowledge distillation to significantly reduce model size and computational cost while maintaining quality in text-to-image synthesis.
   
3. Microsoft and Mistral AI have announced a partnership to boost AI innovation and introduce the ‘Mistral Large’ model on Azure. This collaboration aims to leverage Azure’s powerful computing infrastructure to enhance AI model development and application, promising significant advancements in AI capabilities and services. This partnership is a strategic move to accelerate AI research and practical applications, positioning Azure as a key platform for cutting-edge AI developments.
   
4. Google acknowledged issues with the image generation feature in their Gemini conversational app, leading to a pause in its use. The feature, built on the AI model Imagen 2, encountered problems with generating accurate and appropriate images of people. Google is working on improvements to address these issues and enhance the app’s functionality.
   
5. MIT researchers have developed a machine-learning algorithm to predict drug interactions that may hinder a drug’s effectiveness. This model, based on both tissue models and machine learning, identifies which transporters different drugs use and their potential interactions. For example, it revealed interactions between a common antibiotic and a blood thinner. This approach not only helps in identifying interactions between existing drugs but could also aid in the development of new drugs with improved absorbability or fewer interactions.
   
6. Reddit has entered into a content licensing deal that involves the use of its user-generated content by AI companies. This agreement allows AI firms to access and utilize the vast array of content available on Reddit for training AI models. The deal underscores the growing importance and value of user-generated data in the field of artificial intelligence, especially in enhancing the capabilities and understandings of AI systems. This move by Reddit is indicative of the expanding collaboration between social media platforms and AI technology companies.
   
7. WordPress and Tumblr are reportedly planning to sell user content to AI companies. This move is seen as an effort to capitalize on the vast amount of user-generated content on these platforms, providing AI companies with data that can be used to train and improve AI models. The decision underscores the growing intersection between social media content and the development of artificial intelligence, highlighting the value of user contributions in the evolving landscape of AI research and development.
   
8. Adobe Acrobat introduces the AI Assistant (beta) in Acrobat, allowing users to interact with documents for quick answers and summaries. This new feature enhances productivity by enabling the creation of impactful content through easy engagement with documents. The AI Assistant supports tasks like summarizing key points, generating study guides, and streamlining sales research, all while adhering to Adobe’s AI ethics principles of accountability, responsibility, and transparency.

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’s Tool of the Week: Mistral’s Le Chat

Mistral’s Le Chat, a multilingual conversational AI assistant, represents Mistral AI’s entry into the competitive field of large language models. It is part of their suite of AI models, which includes the flagship model Mistral Large, known for its top-tier reasoning capacities. Le Chat is currently in beta access, showcasing the potential of Mistral AI’s technology.

Le Chat is a significant part of Mistral AI’s offerings, emphasizing the company’s commitment to advancing AI technology and providing versatile tools for AI development. The partnership with Microsoft, announced in February 2024, further expands Mistral AI’s presence in the AI industry, making its language models available on Microsoft’s Azure cloud.

Le Chat’s beta launch underlines Mistral AI’s approach to harnessing AI for practical and innovative applications, contributing to the rapidly evolving landscape of AI and machine learning​

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.