B Cavello

Thousands of often-overlooked workers support every AI system, and their experiences are now receiving overdue attention. In a recent conversation with netzpolitik.org, Joan Kinyua highlights the often-overlooked human cost of AI development. From data labeling to content moderation, these workers play a critical role in training AI systems, yet many operate under precarious conditions with limited protections. Key discussion points that stood out: • Hidden labor: AI is not “magic” — it is built on human effort that remains largely unrecognized. • Accountability gaps: Big Tech is aware of the harms tied to AI production, yet meaningful responsibility and enforcement remain limited. () • Lack of transparency: Consumers and even policymakers struggle to trace how and where AI systems are built, raising questions about ethical sourcing. () • Precarious working conditions: Low pay, short-term contracts, and restrictive NDAs prevent workers from speaking out or organizing. () • Global inequality in AI supply chains: Outsourcing shifts both labor and risk to workers in the Global South, while benefits concentrate elsewhere. () Perhaps most striking is the call to rethink AI itself — not just as a technological breakthrough, but as a system deeply dependent on human labor that deserves recognition, protection, and fair compensation. As AI continues to scale, the question remains: who really pays the price for innovation? 🔗 Read more: https://lnkd.in/ddtUDuCm

38

AI is already shaping decisions about innovation and efficiency, ethical risks and community trust, policies and governance, responsible use, and more. The first findings from #OCLCresearch’s new pulse survey series, Insider Insights, provide a snapshot of how 698 library leaders across the field are thinking about these issues and where they see their institutions headed next. ⤵️ Read the full findings. --- #OCLC #EngagedLibraries #PublicLibrary #PublicLibraries #AcademicLibrary #AcademicLibraries #ResearchLibrary #ResearchLibraries #Library #Libraries #Librarian #Librarians #LibraryLife #LibraryLove

17

Our team was very excited to partner with the Jerome L. Greene Foundation f to build a deliberate, governance-first AI framework with the imperative to protect sensitive data and donor trust, prioritizing clear impact over technological novelty. Read the whole case study here ⤵️ https://lnkd.in/gz5Sv-Ep

2

Last November, COS announced efforts to develop a community-driven strategic plan for ensuring long-term preservation, accessibility, and usability of federally-funded scientific data. Our latest blog post shares an update on the strategic planning process, which is anchored in three core pillars: assessment, foundation, and advocacy. The Committee will be convening in March to begin building strategies to advance these pillars, and is soliciting community input to help inform the plan. 💡 Read more and give your input: https://lnkd.in/ghWuvCjj

3

As a funder, are manual application reviews, fragmented reviewer records, and post-award reporting bottlenecks slowing down your organization? 🏛️💡 While administrative burden affects the whole research ecosystem, grantors bear the heavy lift of verifying applicant data and measuring long-term funding impact. By integrating ORCID into your grant management workflows, you can automate administrative tasks, safeguard research integrity, and give public credit to your peer reviewers. Join us for our upcoming Enabling Value webinar as we feature ProposalCentral by Altum, Inc., an ORCID Certified Service Provider, to show you how ORCID transforms grant administration. What funders will learn: 🔹 Automating workflows: How ProposalCentral + ORCID eliminate manual data entry and automatically add awarded grants and reviewer contributions to researcher profiles. 🔹 Connecting the funding lifecycle: How closing the loop between grant systems and open PIDs builds audit-ready records and reduces reporting fatigue. 🔹 Recognizing reviewer contributions: Practical ways to streamline peer review administration while offering validated recognition to expert reviewers. Close the loop between your grant management system and open PID infrastructure to reduce reporting fatigue and strengthen research integrity. 🚀 = 📅 Date: 30 Sept. 📍 Register here: https://lnkd.in/e859gzsY #ORCID #ProposalCentral #GlobalParticipationFund #ResearchInfrastructure #ResearchFunders #OpenResearch #Grantmaking

9

"Responsible AI that benefits Indigenous communities and conservation must implement #Indigenous data sovereignty principles, and determine..." #Ethics #AI #EthicalAI #LLM #Data https://lnkd.in/eWCJBvba

1

The Aspen Institute Center for Urban Innovation have just released an innovation report. The report highlights four key areas where public institutions can start fostering future-driven innovation. 👉Mindset & Beliefs – Accept that failure is already happening, and use it as a step toward progress. 👉Human Relationships – Build teams grounded in trust, empowerment, and psychological safety. 👉Systems & Processes – Redesign structures to identify, learn from, and act on failure. 👉External Narratives & Ecosystems – Change the public story about what innovation in government means. Sounds simple – but in reality, I often find that these four fundamental elements are rarely in place. Kudos Jennifer Bradley and Dan Vogel & co.

89

Quick experiment: What happens when you feed one of the strongest local government innovation playbooks into Google NotebookLM's new Cinematic video feature? I uploaded the Path to Public Innovation playbook from the Bloomberg Center for Public Innovation at Johns Hopkins University - a go-to resource for city i-teams and innovation offices - and let NotebookLM generate a visual explainer directly from the PDF. No scripting, no editing on my end. https://lnkd.in/g2EC6cwu The playbook itself is excellent and worth reading in full. The question I'm curious about: how well does an AI-generated video preserve the nuance of a resource like this? Especially one built on years of real-world city partnerships? If you've worked with the playbook or are part of the Bloomberg Philanthropies i-team network, I'd genuinely like your take. ⚠️ This is a personal experiment, not produced by or affiliated with Bloomberg Philanthropies. Their communications team does an amazing job. The playbook is their work - I just pointed a new tool at it. #NotebookLM #Cinematic #PublicInnovation #LocalGov #iTeams #GovTech Bloomberg Cities

14

We’re thrilled to see the launch of FRAME (Forum for Real‑World AI Measurement and Evaluation), a new initiative at Virginia State University’s Center for Responsible AI. FRAME is building an evaluation ecosystem to help organizations and communities make informed decisions about AI deployments in line with their operational realities and societal goals. Civitaas helped develop the concepts behind FRAME’s real‑world evaluation approach and is serving in the founding director role in close partnership with VSU’s Center for Responsible AI. Together, we’re working to connect model‑centric evaluation with what actually happens when AI is used in real workflows and institutions. As part of the launch, FRAME has released three new publications that lay out the foundations of this real‑world AI evaluation ecosystem: Real‑World AI Evaluation (LINK) – describes shortfalls of the current evaluation ecosystem and lays out FRAME’s complementary ecosystem to fill those gaps and support stakeholders outside the AI stack. CIRCLE (LINK to https://lnkd.in/eqzhh3_V) – outlines a six‑stage real‑world evaluation lifecycle that follows model behavior to the outcomes that matter for deployment decisions. Context Specification (LINK to https://lnkd.in/e-qDZ6yE) – offers a practical guide for turning deployment priorities into concrete measurement targets and testable questions. To learn more about FRAME, the new publications, and how to get involved, check out the FAQ and resources on the FRAME website: https://lnkd.in/etd_Tssv Huge thanks to the collaborators, community partners, and practitioners who shaped this work so far. Virginia State University Center for Responsible AI https://lnkd.in/efg5QMhp

2

1 Comment