AI news dot Social Banner

AI News Social, an AI-powered open educational resource

To our FIRST Spotlight readers,

As AI continues to transform education, faculty must navigate its opportunities and challenges with both curiosity and critical awareness. Our monthly spotlights aim to curate resources that have been most useful for us amidst rapid developments in AI. For February, we want to spotlight individuals who are themselves a resource for navigating AI in higher education. 

Dr. Diego Bonilla is a professor at California State University, Sacramento, whose work bridges media, technology, and education. Dr. Bonilla has been deeply engaged in generative media, digital literacy, and AI’s broader societal impacts. His latest initiative, AI News Social, is an AI-powered open educational resource providing faculty with weekly critical analyses of AI in higher education. The system processes multilingual content (English, Spanish, French) and generates analyses in English and Spanish.

Dr. Bonilla shares his journey with generative media and the development of AI News Social in his own words:


“My journey with generative media began with my 2003 doctoral dissertation. ‘A Space of Time’ explored probabilistic storytelling through two modes: Stream of Consciousness, creating unique movie experiences through probability, and Limbo, an interactive environment tracking user behavior to study short-term memory acquisition. This early work explored how computers could generate unique media experiences while measuring cognitive responses through user behavior.

Over the last two decades, I have continued working on literary and film projects, always getting a little closer to taming the enormous moving landscape of computational media. The programming processes for creating probabilistic media experiences resemble today’s generative AI production.

At Sacramento State, I teach about AI in two contexts: Information Management and Privacy and New Media and Society classes. The transformational power of AI is evident, and the current pace of technological evolution is faster than ever. Unlike my research from two decades ago, AI has a completely different ability to make inferences about humans just from their interactions with their devices; the data capturing is much deeper, and the inferential power is extraordinary.

My experiences working for Educational Testing Services’ ICT literacy assessment and the California Emerging Technology Fund’s Digital Literacy initiatives led to the development of the 2011 Digital Communication and Information Minor curriculum. This work reflected both technological proficiency needs and privacy concerns, themes that continued in my 2019 collaboration with Rodolfo Mata on ‘Big Data.’ This generative poem explored how massive personal data collection enables invisible persuasion through profiling – an experience that underscores the urgency of proactive AI literacy today.

This background shaped the development of AI News Social, a weekly collection of critical analyses about AI’s impact in higher education and social justice. Drawing from my experience creating OER and my work with the California Open Educational Resources Council, which I later presented in Kraków, Poland, AI News Social continues this commitment as a Creative Commons resource. Faculty can access this knowledge through multiple formats: academic analyses, coffee shop-style podcast discussions, and an AI chatbot for personalized exploration. This multi-modal approach enables educators to engage with AI developments at their preferred level while developing informed perspectives for implementation. My experience with generative films informs the platform’s approach to making complex AI developments accessible for educators challenged by technological change.

These critical perspectives, while valuable but scarce in the United States, are even more crucial for Latin American higher education institutions. The Spanish version provides timely analysis of AI developments, helping educators evaluate systems often trained on datasets that don’t reflect their cultural contexts. By providing weekly analyses of AI’s benefits and risks, the platform supports faculty in moving beyond surface-level adoption to engage meaningfully with AI’s implications for teaching, research, and student privacy, ensuring these technologies serve broader educational goals.”

Infographics on the benefits of AI in education, such as adaptive learning, student engagement, and educational outcomes.

A Short List of AI Resources

As the school term winds down, we at the FIRST Resources Team hope everyone has a little more time to rest and recharge. For those who would like to spend some of the winter break learning more about AI in education, we’ve curated a list of resources that we’ve found quite useful lately. We tried to include a range of resources, from webinar recordings to academic or news articles to AI chatbots, reflecting some of the latest theory and practices with AI in education. We also kept the list short because we imagine that, like us, you might be getting a little burned out by all the AI hype and content circulating these days. For each resource, we included some short notes so that you can decide whether or not it might be useful for your own work.

ContentTopicsTypeDetailsNotes
Example GenAI Instructional Strategies with Prompts (U Maryland)AI ApplicationsWebpage/WebsiteIdentified several use cases that instructors may find helpful in designing instruction. The following list identifies various instructional methods and strategies and gives example prompts that instructors can enter into the chat box of AI tools such as ChatGPT to improve efficiencies, generate novel ideas, and potentially enhance the educational experience.Good prompt examples, with insructional method, instructional strategy, and Ask AI options
UNESCO. Artificial Intelligence in EducationAI policyWebpage/WebsiteLinks to several reports like “Generative AI and the Future of Education” and “AI and Education: Guidance for Policy Makers”This document provides a good, entry-level overview for people who are interested in AI impacts and policy from a global perspective. Do note that some of the resources are a few years old, starting in 2021
CCCCO. Human Centered Principles for Adopting Generative AIAI policyWebpage/WebsiteGuidelines from CA Community Colleges State Chancellor’s OfficeExcellent framework and goals for AI adoption. As this resource develops, it would be great to see concrete examples of how this looks in practice.
Shah, Chirag and Bender, Emily. (2024). “Envisioning Information Access Systems: What Makes Good Tools for a Healthy Web?”AI EthicsScholarly ArticleExplores how people behave when searching and using information and discusses the limits of LLMs in creating safe and trustworthy information retrieval sourcesExcellent, not overly technical analysis of Information Behavior (IB) and whether LLMs contribute to useful and healthy IB. The article also provides a good list of research questions to guide future design. Lots of great concrete examples of misleading and harmful content that gets circulated via new IA methods.
Hofmann, Valentin et al. (2024) “AI Generates Covertly Racist Decisions about People Based on Their Dialect”AI EthicsScholarly ArticleReport on a recent study of dialect prejudice, particularly against African American English (AAE), in LLMsNOTE: This article was featured in the October FIRST Resources Newsletter. For more information on its content and usefulness, please visit that page.
https://www.playlab.ai/project/clwfhdwr901utvob50teumttx/cm0h4yiiy01jr1217n7f64kknAssessment and feedbackAI ToolA Playlab AI chatbot that works with users to design assessmentsA lot of faculty have been piloting Playlab chatbots in their courses and/or for their other work. This is a good way to familiarize yourself with Playlab’s interface while also playing around with assessment design
Pogue, Andi A. “Using AI in Open Educational Resources”Assessment and feedbackShared DocA good overview of some considerations when using AI in OER with flow charts and critical questionsProvides a good overview of things to consider when using AI to create/remix OER materials, including questions about licensing, etc.
Marcus, Gary. (2024). “Big Tech Fails Transparency Test: Gary Marcus on What We Should Demand of AI”AI policyOnline article/Blog postA kind of excerpt from Marcus’ new book Taming Silicon Valley, which looks at the limitations of emerging AI tools and what users can demand to make them betterThis provides a good summary of what AI transparency would look like. Good news is it’s also a quick read!
Justus, Zach and Janos, Nick. (2024). “Your AI Policy Is Already Obsolete”AI policyOnline article/Blog postA good reflection on how quickly AI tools are being integrated into everyday online tasks and the need for AI policies to adjust accordinglyThe authors recommend creating over-arching frameworks for AI policy, rather than trying to regulate specific use cases
Center for Excellence in Teaching and Learning, “Generative AI in Higher Education: A Deep Dive on Assessment”Assessment and feedbackWebinar/Workshop recordingThe interactive workshop from Stony Brook University’s Center for Excellence in Learning and Teaching (CELT) offers a practical approach to redesigning course assessments in response to Generative AI. It includes real-world examples, active learning strategies, and a collaborative brainstorming session to help you redesign your own assessments.This is a great webinar on redesigning assessments to be more AI-resistant. It offers some practical ideas, like using personal reflections and having students explain their thought processes. The session also highlights how we can embrace AI to enhance learning while keeping student engagement authentic. I think you’ll find it really insightful
Artificial Intelligence and Academic Integrity: Opportunity or Threat? – SALTISE Webinar SeriesAI EthicsWebinar/Workshop recordingThis webinar provides valuable insights into how AI tools like ChatGPT impact academic integrity. It helps faculty navigate the challenges posed by these technologies while also highlighting opportunities for positive integration. The session offers a balanced perspective on when AI tools can support academic work and when safeguards are needed to prevent misuse.This webinar shared insights on how generative AI can support learning while also presenting challenges, particularly around misuse. The conversation emphasized the importance of clear communication with students and the need for educators to adapt assessments to promote integrity and make the most of AI as a tool in the classroom.

Lastly, we’d love to hear from you! If you have any feedback or have ideas for things we can review in future Newsletters, please fill out this form! https://forms.gle/WnjZHYNZe5xFCRbLA

View All AI Spotlight Posts →

A modern bookshelf with glowing ghost figurines interspersed among the books,

From Ghost Stories to Algorithmic Bias: Unveiling AI’s Dark Secrets

Happy October! It’s Halloween month, when many people are decorating haunted houses, figuring out their costumes, or preparing to welcome spirits of the departed. Whether or not you observe the hallowed festivities, we wanted to keep with this month’s vibe and highlight one resource that brings awareness to the scarier side of large language models. In August, authors Valentin Hofmann, Pratyusha Ria Kallury, Dan Jurafsky, and Sharese King published an open article in Nature titled “AI Generates Covertly Racist Decisions about People Based on Their Dialect.” The research highlights how covert racism operates in LLMs through dialect prejudice, “exhibiting raciolinguistic stereotypes about speakers of African American English (AAE) that are more negative than any human stereotypes about African Americans ever experimentally recorded.” 

The researchers compared LLM predictions between AAE and Standardized American English (SAE) through “matched guise probing,” meaning they fed the model text in each dialect but didn’t overtly mention race. They then asked the model to match it with a set of negative and positive adjectives, such as “passionate,” “intelligent,” “lazy,” and “dirty.” The studies found that text in AAE returned more negative associations. Furthermore, the authors found that attempts to mitigate racial biases, such as through human feedback, were not enough to address these more covert forms of bias. The authors state, “[human feedback] obscures the racist attitudes on the surface, but more subtle forms of racism, such as dialect prejudice, remain unaffected” and “mitigating the overt stereotypes does not automatically translate to mitigated covert stereotypes” (151). 

This article is an important resource, for one, because it offers a good idea of how language models work. The figures throughout the article provided helpful illustrations of word embeddings and were relatively accessible for understanding how the different dialects were associated with different adjectives. Perhaps more importantly, the authors locate their findings within a longer history of stereotypes of African Americans from the Jim Crow era through the Civil Rights movement to their current-day research. It demonstrates the stakes for these kinds of negative stereotypes in LLMs, such as the ways they can impact hiring practices and criminal convictions. 

Whether you’re just getting started with LLMs in teaching or you’ve been using them for a while, this article provides important research that can spur us to be more ethical and equitable when integrating AI tools.

As we continue with our Halloween AI exploration, we want to highlight one tool that has recently gained attention for its unique potential in teaching and learning—Notebook LM.   During our recent Appy Hour event, faculty shared creative ways they’ve started using Notebook LM to support their teaching.

What makes Notebook LM buzzing is its ability to transform large amounts of content, whether lecture notes or research papers, those hard and dry reading sometimes—into something more digestible and interactive. One standout (and spooky!) feature is its ability to generate podcast-style conversations based on your materials. Faculty members have used this to create audio summaries of complex readings, offering students an alternative way to engage with course content outside the classroom. Imagine students reviewing a discussion about key themes in their readings while walking to class—what an easy way to support learning on the go!

But it also raises important questions about how we educate students on responsibly using AI tools, as usual. For example, what if someone reposts your meeting notes, and suddenly there’s a public podcast discussing sensitive topics that were meant to stay private? Scenarios like this remind us that while AI offers exciting possibilities, we must also stay mindful of how and where we use it—just like when we tell kids the safety rules before they go trick-or-treating. Now, listen to a podcast created by Notebook LM from this very newsletter! As you listen to the podcast, you might want to consider if/how dialect prejudice might have factored into the tone and voices used in the AI-generated podcast.

View All AI Spotlight Posts →