Janna Quitney Anderson

🎧 The next episode of “AI, Unscripted” is Monday, Sept. 8, featuring UMCP’s Ben Shaw on AI & Information Literacy! In the upcoming “AI, Unscripted episode,” @Mary Crowley-Farrell and @Jennifer Potter talk with @Ben Shaw from @University of Maryland, College Park, about the supports Ben and his team have built to help students and faculty gain information literacy skills around Generative AI. Why this matters: "In early 2023, about half of the requests we received at the library were for non-existent materials," Ben reveals, highlighting the urgent need for AI guidance in higher education. Key Insights from Ben: -The strategic decision-making process behind their AI & Information Literacy module -A walkthrough of the module’s content and implementation -Real feedback from faculty and students using these resources -How to access and embed this openly licensed module in your own courses Want to view the module or download it for your course? Visit the @UMD TLTC page at https://lnkd.in/eE5w4w25. The episode will be available on Apple Podcasts at https://lnkd.in/egvmnUx8 and Spotify at https://lnkd.in/eKdWRa_g. Or you can catch it on the Kirwan Center's website at https://lnkd.in/eV_9ZUpe! “AI, Unscripted” is a limited series from @University of Maryland, Baltimore's Moving the Needle podcast, co-sponsored by @USM Kirwan Center for Academic Innovation, @CUSF, and @MarylandOnline. What's your next step in becoming AI-confident? Share using #AIUnscripted or email us at aiunscripted@usmd.edu. And please don't hesitate to pass this information on to your colleagues, too! We’re hoping these conversations will inspire college and university faculty across the state to take their next step, however small, in exploring AI's potential in their own teaching practice.

18

2 Comments

Had a great discussion with journalist Neenma Ebeledike on tech fatigue, especially the new AI fatigue, last week. As small businesses navigate the AI era, leadership, strategic planning, and clear expectation communication are key for successful adoption, more in this 👉🏼https://lnkd.in/gFNJ7gAQ

14

1 Comment

We're going to need journalists to stop talking about synthetic text extruding machines as if they have *thoughts* or *stances* that they are *trying* to *communicate*. ChatGPT can't *admit* anything, nor *self-report*. Gah. https://lnkd.in/eMunSQfh

601

60 Comments

We are officially live!! 💙 The website for the UCSB Media Industries and AI Research Initiative is now up, and we are thrilled to join this growing global community of twenty one researchers from fifteen institutions across seven countries and three continents. It is an honor to be part of a cohort that brings together a wide range of disciplines and methodologies to examine the global integration of AI through questions of labor, post production, media distribution, audience engagement, intellectual property, and policy. I am excited to contribute our project Synthetic Voices, Precarious Labor: AI and Voice Acting in the Video Game Industry, which explores how AI is reshaping creative labor, performance, and the future of voice acting in gaming. I am deeply grateful to collaborate with Kate Fortmueller and to work alongside such an inspiring network of scholars and artists whose projects form essential components of this collective endeavor. Here is to building new avenues for collaboration, comparative inquiry, and future thinking together. Georgia State University Georgia State University College of Arts & Sciences WGSS@GSU 💙

12

A thoughtful (paraphrased) comment from Mohammad Hosseini that has particularly stuck with me following National Information Standards Organization's webinar on Research Integrity earlier today: The use of generative AI is a double edged sword. On one hand it can help save time during the research process, but on the other hand it adds time to the process as every claim, quote, and citation put out by generative AI tools neeeds to be double checked against the original source for accuracy. This verification process often costs more time than the initial prompt saved. Responsible and transparent use of generative AI can greatly benefit the research process, but ultimately it is the researchers' responsibility to ensure that the output is correct, reliable, and free from hallucinations and incorrect interpretations of the source data.

22

1 Comment

Appreciated the opportunity to join today’s Business Buzz conversation live with Ryan Frazier on KUSA-TV, 9NEWS. AI-driven workforce changes are complex, local, and deeply human. What matters most is how leaders respond, with intention, transparency, and innovation rather than fear. Important conversations like this help move the discussion from headlines to insight. #AI #FutureOfWork #Leadership #BusinessAndTechnology #PublicEngagement #WomenInTech #LiveTv

19

Could not agree more with this insightful piece about AI chatbots from Michal Luria, Ph.D. in Tech Policy Press. We don't need them to sound human. But we do need them to stop manipulating our human instincts for connection. Not only does it waste time and untold natural resources powering these chatty attachment algorithms, but these features are leading to real human harms that are only now beginning to come into view.

4

Back in May we ran a workshop at Northwestern University on how we might develop a journalism-specific benchmark for LLMs. The idea was to be able to support journalists so they could better compare this versus that model, but for news-specific tasks, and in terms of the values important to journalists. Today, we're publishing our attempt to move the needle. The basic tension is that generic benchmarks aren't that useful for decision-making since there are roughly a gazillion different ways you can encounter a task like "information extraction" in the course of doing journalism. Our approach is premised on putting the power in the hands of journalists to edit and adapt a set of templates. We create the structure (and a few different examples) and you can take that and adapt it for your specific context, task scenario, and data. It's all implemented in a set of Google Colab notebooks that you can copy, adapt, and run. Here's our blog post on GAIN explaining it. We're calling it The Journalism Benchmark Cookbook: https://lnkd.in/gQcPN8RS What do you think? cc Charlotte Li Jeremy Gilbert

80

6 Comments

Maintaining the Scholarly Record Many academics understand the value of DOIs (digital object identifiers) as an essential scholarly infrastructure. Journals submit their metadata to Crossref, so the DOIs know where to link to the articles. But what happens when a journal moves to a new website? While the DOIs never change, they need to be updated, so they take people to the right webpage. That’s where maintaining DOIs comes in. Updating more than 900 DOIs from 87 volumes (years) of a journal, each with a new URL, takes systematic thinking and fine-detail accuracy. One wrong digit can mean people cannot find an article. This helps ensure researchers can find, cite, and share these articles. It may have only taken a few hours of focused work, but it feels good to complete this project for my client.

3

1 Comment

Infuriating on so many levels. How Grammarly thought it was okay to use mine and many others' names in their product is so incredibly disrespectful and exploitative. They've charged money for a feature with implicit suggestions that myself and others are okay with this. That my "expert" review could lead some student astray seems not to matter at all to Grammarly, and I'm sure that they're argue that it's on an AI likeness "inspired by" my work. If you're so inspired by my work, consider asking me directly. In an era where we pay so much more attention to name, image, and likeness ... I'm curious how this plays out. Maybe I'm just John Henry to the Iron Horse here, but when folks ask me why I'm skeptical of rapid and unchecked AI deployment, it's not out of fear. It's out of bonafide fraud.

42

14 Comments

clear cut illustration from Alonso Guzmán Barone: A searchable online repository can restore trust in federally funded basic science: Check out U.S. Public Research Benefits: https://lnkd.in/dHWuqwU7 via The Transmitter https://lnkd.in/dC4xV6cq

37

3 Comments

I am excited to share the release of a brand new audio program called Science Engaged, produced in partnership with the Civic Science Media Lab, that explores why and how scientists partner with diverse communities. In the program, I speak with scientists who have worked with local communities on a variety of participatory science projects. I ask them why they think public participation in science is important, what results these collaborations produce, and how to do it well. This program and episode is made possible by support from Schmidt Sciences and the UC Santa Cruz Science Communication Program Catalyst Grant. Special thanks to my co-producer Fanuel J. Muindi

49

2 Comments

I'm a fan of Eric Seufert's Mobile Dev Memo Podcast. https://lnkd.in/g5qD9P7u. The latest episode is terrific, with Carl Mela and Ross-boy Link talking about measuring the long term effects of marketing spending on sales and profitability. What pros these two guys are, with deep experience with implementing these models. Many, many insights, but the core ideas were teed up by Carl at the outset about what most companies get wrong about brand measurement. 1. Brands take years to build, requiring some change in their long term preferences. 2. Too often brand strategy is based on what is easily measured rather than letting strategy drive measurement. Advertising builds brands, but its short and long term effects are small and thus hard to measure. This led to shifting ad budgets to direct response, as if a click is the key KPI. 3. Firms are tied up in targeting rather than reach, on the idea of focussing advertising on the customer most likely to respond. Growth however comes from growing the franchise. 4. Finally firms attempting to measure brand effects through Media Mix Models often overlook the effects on brand of other marketing mix elements. Carl has shown that product and distribution swamp advertising in effects on sales and profitability. So think marketing mix modeling rather than media mix modeling. Great episode. Even the most sophisticated listener is going to pick up some new insights.

19

5 Comments