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, IBM unveiled Granite 3.0, a suite of enterprise AI models focused on secure data processing. MIT introduced an innovative method, Heterogeneous Pretrained Transformers, to train general-purpose robots more flexibly. Physical Intelligence showcased an AI model with the ability to do a wide range of household chores. Anthropic launched Claude 3.5 Haiku, a model designed with enhanced speed and reasoning capabilities. Meanwhile, Stanford’s Michal Kosinski argued that GPT-4 may demonstrate a skill similar to of “theory of mind,” sparking new discussions on AI’s potential for social cognition. Lastly, an article explores whether AI models are genuinely capable of reason or simply recognize patterns.
Here’s this week’s news:
IBM Granite 3.0 Unveiled
IBM has introduced the third generation of its Granite AI models, designed for enterprise tasks such as cybersecurity, text analysis, and content generation. The Granite 3.0 models, available in various configurations, prioritize performance, safety, and open-source availability under the Apache 2.0 license. Read more.
MIT’s Efficient Training Method for General-Purpose Robots
MIT researchers have developed a new approach to train general-purpose robots more effectively by integrating diverse data from varied sources, inspired by large language model techniques. The architecture, named Heterogeneous Pretrained Transformers (HPT), consolidates data from different modalities and domains into a unified system, significantly enhancing adaptability across tasks while requiring far less task-specific data for each new scenario. Read more.
Physical Intelligence’s Advanced Home Robot Capabilities
Physical Intelligence, a new startup, has developed a general-purpose AI model for home robots, capable of performing tasks like laundry folding, table cleaning, and box building. By training its model on extensive amounts of robotic data, the company aims to bring the adaptability of AI models like ChatGPT to robots handling tasks in the physical world. Read more.
Claude 3.5 Haiku’s Advanced Capabilities
Anthropic’s Claude 3.5 Haiku, the latest model in its AI lineup, offers enhanced speed and reasoning skills suitable for tasks such as coding, data labeling, and real-time content moderation. Designed for efficiency, it supports rapid user-facing interactions, and is ideal for customer service, development workflows, and large-scale data processing. This model also upholds Anthropic’s high safety standards, showing improved handling of sensitive content and complex reasoning tasks. Read more.
Stanford Researcher Explores AI’s Potential to Understand Humans
Stanford researcher Michal Kosinski claims that advanced language models like GPT-4 may have developed an ability close to “theory of mind” (ToM), the cognitive ability to interpret other people’s thought processes—a skill once thought to be unique to humans. In recent tests, GPT-4 performed comparably to young children on tasks requiring ToM, suggesting an unintended emergence of social-cognitive skills as models advance in language understanding. Although some researchers argue this capability may simply reflect memorized patterns, Kosinski believes such progress hints at AI’s growing potential to understand, predict, and even influence human behavior with remarkable precision. Read more.
Can AI Truly Reason, or Is It Just Pattern Recognition?
This article examines whether large language models exhibit genuine reasoning or merely advanced pattern recognition. While LLMs perform impressively on benchmarks, critics argue that their capabilities may not reflect true reasoning, as slight changes in prompts can disrupt their accuracy, revealing a reliance on statistical patterns. The article suggests that AI’s success in “reasoning” tests may be more about matching patterns rather than understanding concepts, prompting calls for new evaluation methods to test AI’s reasoning potential more rigorously. Read more.
Tools of the Week

Tool of the Week: ChatGPT Search
What is ChatGPT Search?
ChatGPT Search is a newly integrated feature in OpenAI’s ChatGPT that enables users to access up-to-date web-based answers directly within the chat interface. Launched for Plus and Team users, this feature allows ChatGPT to respond to questions using real-time information by sourcing and displaying relevant online content.
How is it used?
Users can access ChatGPT Search through a simple search icon within the chat or automatically when ChatGPT detects a query needing web information. It offers detailed responses enhanced by links to original sources, making it easy for users to verify or explore information further on topics like news, weather, sports, and finance.
What is it used for?
This tool is ideal for retrieving fast, trustworthy insights on current events, stock prices, weather updates, and more, blending the convenience of conversational AI with real-time data. With partnerships across various media outlets, ChatGPT Search enhances information accessibility for users looking for up-to-date information from reliable sources.
Learn more about ChatGPT Search here.
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 […]