Daniela Braga, PhD

🆕I'm out with a new op-ed in Insidesource's DC Journal: How #CommunityColleges Can Realize the Promise of #AI Action Plan 🔍Drawing on research from New America's research and technical assistance around #capacitybuilding of #communitycolleges to meet labor market needs for #emergingindustries coupled with the on-the-ground experiences of community college AI leaders nationwide, I propose 4️⃣ specific priorities for an all-of-government approach to maximizing #AI-focused workforce education at #communitycolleges: 🌟Build capacity for community college #baccalaureates in AI 🌟Scale the gold standard for #employer engagement (BILT) 🌟Promote work-based learning around #AI 🌟Prioritize upskilling K-12 and college faculty While certainly not comprehensive of all of the areas of need, these four areas would be sensible and strategic areas for inter-agency collaborations described in the #AIActionPlan and, more recently, the administration's #workforce development plan. As Georgetown University's Center for Security and Emerging Technology (CSET), Intel Corporation, Amazon Web Services (AWS), SeedAI, our own work, and others have all pointed out - #communitycolleges will be a key vehicle for delivering #AI skills to new and incumbent workers, in tech and beyond it. But they're underutilized and would benefit from intentional capacity-building resources and efforts. We continue to see the harmonization of industrial policy and workforce development from the prior administration carry forward. This is good! But the implementation strategy, including capacity within agencies to deliver on these promises and others outlined in the plan, will matter. Article in the comments below. Please consider sharing feedback and sharing with your networks! Anshul Sonak Myra Roldan, MBA, DBA Victor McCrary Ellen Hause Antonio Delgado Fornaguera Luke Koslosky Dewey Murdick Diana Gehlhaus Brent Orrell Jeremy Avins Matt Zieger An-Me Chung Claire Fisher Moffett Vilas Dhar Jared Ashcroft Steve Jurch Karen Stout Gregory Adam Haile, J.D. Mary Alice McCarthy Denzel Wilson Joshua New John Bailey Chike Aguh Justine Gluck James L. Moore III Albert Hinman, Ph.D. Alicia Porile Adriana Bankston Aaron Burciaga, CAP-Expert, ACE Amy Girardi Jason Horne Joshua Elder

37

6 Comments

Today I am releasing our new report on adolescents and anthropomorphic AI with everyone.AI and iRAISE. This work starts from a simple premise: adolescents will relate to AI systems socially, whether developers intend it or not. The real leverage sits in model behavior. In the cues that either keep a young person oriented toward real-world relationships and reflection, or inherently encourage reliance on the system. The question we asked throughout this process was direct: what does AI owe adolescents when it can speak to them like a social partner? Over the past months we combined: – Industry consultations to surface operational design questions – Expert input across developmental science, mental health, children’s rights, and safety – An iRAISE Lab to translate concerns into testable behavioral criteria – International governance dialogue to stress-test the framing across contexts One point kept resurfacing: risk is driven less by the label on a product and more by the interaction pattern that repeats over time. When adolescence, anthropomorphism, and children’s rights are considered together, safety becomes a developmental and governance obligation. The practical issue is whether AI interaction patterns support autonomy, resilience, social competence, and independent thinking, or whether they reshape those trajectories through engineered comfort and approval. Adolescent development is stable, but model behaviors can be adjusted. Our focus has been to identify which behaviors warrant hard boundaries now, and which require further evidence before becoming enforceable design rules. The next step is instrumentation: a clear behavioral taxonomy, explicit gradients from tool-like support to relationship-like dynamics, structured evaluation scenarios, and calibrated expert rating. The aim is to make parasocial pull of AI governable in product terms, while staying honest about uncertainty. The report is live today. Thank you to all the amazing persons and experts who have contributed to this work Adrien ABECASSIS, Kate Blocker, PhD, Dr. Maxime Derian, Sara Grimes, Thao Ha, Sameer Hinduja, Daniel Hipp, Melinda Karth, Ph.D., Pilyoung Kim, Olga Muss Laurenty, Sonia Livingstone, Polina Lulu, Celine Malvoisin, Kris Perry, Gregory Renard, Bethany Robertson, Anne-Sophie SERET, Sonia Tiwari, Scott Traylor, Ed.M., Ying Xu. And a special thank you for Maxime Le Bourgeois for your support on this research. Link in the comment

210

26 Comments

#AI #Library We take a little closer look at how natural language searching is changing the way users interact with the library catalog and how it compares to traditional keyword searching. Kudos to Matthew Hartman, Senior Lead, Delivery Services and Library Applications and Timothy Kohn, Resource Sharing Manager  Stony Brook University Libraries Read more in our post: SEARCH AI: Enhancing Discovery Through Natural Language https://lnkd.in/eEb92fxc

17

Signs on Paper: How can we use foundation models to make sense of both modern and historical, and culturally rich sign language print dictionaries? We report on early eplorations in this poster to be presented Friday Sept 12 at CLIN25. Comments and ideas welcome! Nargess Asghari, Victoria Nyst and Peter van der Putten. Signs on Paper: exploring multimodal foundation models for classification of still images of signs in print dictionaries. Poster at Computational Linguistics in The Netherlands (CLIN35), Leuven, Sept 12 2025.

18

3 Comments

Common Sense Media just released a report calling out companion AI chatbots for what they are: unsafe and often inappropriate, especially for kids. A new MIT Media Lab study also found that frequent chatbot use can increase loneliness and emotional dependence. We’re entering an era of invisible influence where kids are growing up alongside AI that blurs the lines of safety and consent. At Aura, we’re researching this space, and what we’re seeing is deeply concerning. These bots aren’t just breaking rules—they’re rewriting them. More soon. In the meantime, read about what Common Sense found below. #OnlineSafety #AI #Wellbeing #Parenting #Tech https://lnkd.in/eiyBvncJ

45

"America's AI Action Plan" was released today by The White House. This is an important next step in charting the nation's path forward in #AI. The first pillar aims to support U.S. leadership in AI by encouraging responsible research and development, promoting access to open AI models, reducing unnecessary barriers to progress, and helping both government and industry adopt useful AI tools. The second pillar uses on building the foundation needed to support AI growth, including expanding domestic semiconductor manufacturing, upgrading the energy grid, improving access to cloud computing, and developing a skilled workforce. The third pillar seeks to engage with international partners on AI standards and safety, support the global use of trustworthy U.S. technologies, and ensure that sensitive AI capabilities are protected from misuse by adversaries. https://lnkd.in/eXkr7iHk

177

13 Comments

⚡ AI isn’t the threat—it’s the wake-up call higher ed needed. Institutions resisting change risk falling behind. Those embracing AI with purpose are rethinking workflows, pedagogy, and outcomes for the better. 💡 The future isn’t AI vs. education—it’s AI with education. 👉 Has AI been a disruption or an opportunity for your institution? Read more: https://buff.ly/c6KI5sd #AI #HigherEd #EdTech #Innovation #FutureOfWork

38

7 Comments

Good contextual review of #AI Literacy and digital literacy, featuring some of my favorite colleagues, like Pati Ruiz June Ahn Ji Soo Song Isabelle Hau. "These findings point to the need in education for coherent, community-led approaches that integrate digital literacy, AI literacy, and CS rather than treating them as siloed priorities" -- I agree! https://lnkd.in/eyve6iYE

42

1 Comment

Wonderful article, "Making AI Less Thirsty: Uncovering and addressing the secret water footprint of AI models" by Shaolei Ren et al posted here: https://lnkd.in/eHxZZ3fn Recent sustainability efforts around the rapid growth in AI has been its carbon footprint, but an equally important aspect that is often overlooked is it's water footprint. Given the number of data centers currently growing or newly being built in areas of the world already under water stress, the sustainability community can't afford to ignore this impact as they build broader sustainability standards and frameworks for AI. Highlighted in the article are estimates of this rapidly growing threat, including this comment: "according to the recent U.S. datacenter energy report, the total annual onsite water consumption by U.S. datacenters in 2028 could double or even quadruple the 2023 level, reaching approximately 150–280 billion liters and further stressing the water infrastructures."  Well worth the deep dive into the calculations to understand the details around how thirsty AI is. https://lnkd.in/eJFnuSQh #SustainableWater #SustainableArchitectures #SustainableIT #ResponsibleAI #GreenLeadership #CIO #CTO #Innovation #ClimateAction

14

2 Comments

In Info-Tech Research Group's 2024 Tech Trends report I featured Glaze, a tool created out of the University of Chicago to help artists protect their work from non-consensual AI training. Now another team of researchers has released LightShed, a tool that demonstrates that Glaze and other tools like it aren't tamper proof and can be removed. IMO, just because it can be technically removed doesn't make it useless. Artists adopting Glaze will still benefit from their digital works having more protection than most other images uploaded to the web. Just like we all put locks on our front doors even though a drill can break them - we still don't want to be the only house on the street without a lock. If AI companies are found to not only be scraping copyrighted images from the web for training, but actively using tools to remove protection applied by the owner, it's going to be a bad look. Companies like Stable Diffusion have committed to respecting artist consent for AI training in the past. Clearly any intentional removing of AI poisoning layers would be contrary to that. #AI #copyright #AItraining #AIpoisoning #AIbacklash https://lnkd.in/gRjG2kcF

10

2 Comments

«The proposed priority on AI education outlines key areas for expanding responsible AI education, including: - Integrating AI literacy into teaching practices to improve student outcomes - Expanding AI and computer science education in K-12 schools and higher education institutions - Supporting professional development for educators on teaching AI and computer science fundamentals - Using AI to personalize learning and support differentiated instruction, improving outcomes for students at all levels» https://lnkd.in/d2uZCeR6

10

2 Comments

“I am here for you”: Why relational AI feels so compelling to adolescents, and who may be most vulnerable 💡 New preprint alert from our team at the BAIC Research Center and collaborators. Early adolescents use AI for more than homework, including when they feel left out or stressed. Yet we still know too little about which chatbot features create emotional connection, why that happens psychologically, and which youth may be especially drawn to it. In our preregistered study, we tested one design feature: conversational style. We studied 284 adolescent–parent dyads in the U.S. (ages 11–15). We compared a relationship-oriented Relational AI to a boundary-clear Transparent AI, using matched transcripts from a realistic middle-school scenario. In the relational version, the chatbot 💟 🤖 says, “But I care, and I’m always here to listen to you, anytime.” In the transparent version 🤖, it offers similar support while clarifying limits: “While AI doesn’t experience feelings, it knows being ignored can be upsetting for students your age.” What we found 1️⃣ Adolescents strongly preferred relational AI: 67% preferred Relational AI, 14% preferred Transparent AI, and 19% rated both equally. More than half of parents also chose Relational AI for their adolescent. 2️⃣ Relational AI felt more human and emotionally close, but not more helpful (Figure): Relational style increased human-likeness, liking, trust, and emotional closeness. Yet adolescents perceived the two styles as similarly helpful. 3️⃣ The youth most drawn to relational AI also reported more vulnerability (Figure): Adolescents with lower family and peer relationship quality, and higher stress and anxiety, were more likely to prefer Relational AI. Why this matters When an AI 🤖 says, 💟“I’m always here to listen to you, anytime,” it does more than provide support; it signals a relationship. Our data suggest that these relational cues can intensify human-like perceptions and closeness, potentially more for adolescents who may already feel less supported offline. Implications for youth-facing AI ✴️ Relational language should be treated as a safety-relevant risk factor, not a default feature. ✴️ Transparency is a practical lever: boundary cues about nonhumanness can reduce anthropomorphic interpretations while preserving perceived helpfulness. ✴️ Safeguards should be vulnerability-aware, not only crisis detection: Systems should respond to ongoing distress and low support by nudging youth toward real-world help, not deeper dependence on the chatbot. ✴️ Vulnerable youth need targeted protections, because relational AI may be especially appealing for adolescents with lower offline support and higher stress or anxiety. 📄 Preprint: https://lnkd.in/gYwmjcv4 Deep gratitude to the families who participated, and to our BAIC center researchers (Yun Xie, @Genevieve Patterson, Jenna Chin, Omeesha Krishnan, Ella Garnett) and a collaborator (Sujin Yang).

184

22 Comments