Episódios

  • Aral Surmeli on HERA’s Frontline Experimentation with AI

    Aral Surmeli on HERA’s Frontline Experimentation with AI
    Jul 6 2025
    How can humanitarian organizations harness the power of AI to deliver critical healthcare when the technology is advancing faster than ever, but essential aid funding is being cut? Aral Sürmeli, Executive Director at HERA Digital Health, discusses this question with Humanitarian AI Today guest host and podcast producer Brent Phillips, sharing insights and lessons from his team's frontline experiments with AI. Aral and Brent also discuss the AI for Good Summit, explainable AI and AI research.
    Exibir mais Exibir menos
    31 minutos
  • A New Ethical AI Adoption Toolkit for Humanitarian Actors

    A New Ethical AI Adoption Toolkit for Humanitarian Actors
    Jun 29 2025
    On this special short episode of Humanitarian AI Today, guest host Brent Phillips sits down with Tigmanshu Bhatnagar, a lecturer at University College London (UCL), and Hamdan Albishi, a UCL MSc student in AI for Sustainable Development. Tigmanshu and Hamdan discuss a toolkit they are developing, designed to empower non-technical humanitarian actors to build their own ethical AI projects. They walk through the toolkit's four-phase process—Reflection, Scoping, Feasibility Assessment, and Development—which guides users from an initial idea to a simulated, ethically-sound AI project without needing deep technical expertise. Toolkit users define a problem, identify beneficiaries, and consider potential unintended harm. The tool presents existing use cases and projects in the same problem area to educate the user. The toolkit helps users assess project feasibility based on resources and regulations. It can also suggest publicly available humanitarian datasets and helps check them for completeness and bias to avoid unintended harm. The tool suggests appropriate technical solutions, generates a project with embedded ethical guardrails, and runs it in a simulated environment to validate its accuracy and impact before real-world deployment This initiative emerged from a UK Humanitarian Innovation Hub (UKHIH) and Elrha-funded project, which found that humanitarian organizations, despite their commitment, faced a steep learning curve in creating tangible AI solutions. The toolkit addresses AI adoption challenges and aims to help humanitarian actors develop responsible AI projects for users, regardless of their technical background.
    Exibir mais Exibir menos
    29 minutos
  • André Heller on Piloting Agentic AI and Client-facing Applications at Signpost

    André Heller on Piloting Agentic AI and Client-facing Applications at Signpost
    Jun 23 2025
    André Heller, Director of Signpost, led by the International Rescue Committee, updates Humanitarian AI Today podcast producer Brent Phillips on Signpost’s latest AI publications and research findings. André and Brent discuss a new Signpost paper covering their work piloting agentic AI and client-facing applications and touch on funding-connected complications caused by the aid funding crisis. This short episode was recorded to lay groundwork prior to recording a more formal interview featuring André and Mala Kumar, Head of Impact with Humane Intelligence, covering Signpost’s latest work in greater detail.
    Exibir mais Exibir menos
    18 minutos
  • Till Trojer and Rory Crew on Cash Assistance and AI Strategy: Insights from HEKS and CALP

    Till Trojer and Rory Crew on Cash Assistance and AI Strategy: Insights from HEKS and CALP
    Jun 7 2025
    Till Trojer, AI Officer at HEKS/EPER (Swiss Church Aid), joins Humanitarian AI Today guest host Rory Crew, Technical Advisor on Data and Digitalization with the CALP Network which works on increasing the quality, quantity, and impact of humanitarian cash and voucher assistance, to discuss how mid-sized aid organizations are developing and implementing their AI strategies, and the impact of the aid funding crisis on the sector. Till's role at HEKS/EPER involves helping the organization use AI responsibly, building trust, and ensuring ethical standards by understanding the context and listening to the concerns and needs of teams. Till provided insights into HEKS/EPER’s AI approach, explaining how AI is integrated into their broader digital transformation strategy rather than being treated as a separate entity. He also shared his views on AI as a catalyst for deeper discussions about existing systems, infrastructure, and data hosting dependencies, prompting a critical re-evaluation of reliance on big tech companies. Rory highlighted the timely nature of the conversation, as both cash assistance and AI offer the potential for greater efficiency—a desperate need for the sector amidst calls to do more with less. Both guests emphasized the importance of ensuring that AI solutions are grounded in the communities they serve and avoid the "graveyard of humanitarian pilots" by planning for sustainability and community buy-in from the outset. Ultimately, their vision for futuristic AI centers on systems that prioritize ethics, accountability, and empowerment, refusing unethical tasks and ensuring that communities remain in the driving seat of their own development.
    Exibir mais Exibir menos
    58 minutos
  • A Critical Discussion on Assessing Humanitarian AI Use Cases

    A Critical Discussion on Assessing Humanitarian AI Use Cases
    May 25 2025
    On this special Humanitarian AI Today episode, focusing on assessing evidence on the effectiveness of humanitarian AI use cases, Alexandra Pittman, CEO of Impact Mapper, Suzy Madigan, Responsible AI Lead with Care International, Gary Forster, CEO of Publish What You Fund, and Linda Raftree and Quito Tsui from the MERL Tech initiative discuss challenges associated with assessing the effectiveness of humanitarian aid activities and activities incorporating applications of artificial intelligence. The discussion touches on challenges associated with collecting, mapping, measuring and assessing data on humanitarian needs and aid activities and activities incorporating uses of AI, and on fundamental questions surrounding crisis contexts, stakeholder engagement, operations, reporting, data transparency, uses of technology and AI, and limits on what the sector can actually deliver around impact. This podcast panel discussion was originally recorded in 2024 to help inform an accompanying discussion focusing on real-world AI use cases and their impact on humanitarian action co-produced by Humanitarian AI Today, the UK Humanitarian Innovation Hub and Elrha for a special six-part panel discussion series critically examining different aspects of humanitarian applications of artificial intelligence.
    Exibir mais Exibir menos
    48 minutos
  • Valentine Pistorozzi on Advances in AI and Boldcode’s Reporter AI Project

    Valentine Pistorozzi on Advances in AI and Boldcode’s Reporter AI Project
    May 14 2025
    In this short Humanitarian AI Today podcast episode Valentine Pistorozzi speaks with Brent Phillips about Boldcode, new projects in the pipeline, and changes in the way that AI applications are being built. Brent and Valentine speak in detail about Boldcode’s https://reporterai.org/ project and about technical aspects of humanitarian AI applications. This interview was recorded as a test in November 2024 to experiment with a short 15 minute interview format for a special series of new Humanitarian AI Today podcast episodes to be published on Mondays.
    Exibir mais Exibir menos
    29 minutos
  • Luke Marsden from HelixML on the Past, Present, and Future of Generative AI and MLOps

    Luke Marsden from HelixML on the Past, Present, and Future of Generative AI and MLOps
    May 5 2025
    Luke Marsden, CEO of HelixML, speaks with Humanitarian AI Today podcast producer Brent Phillips about how generative AI has evolved since early language models like BERT were introduced and applications like ChatGPT captured widespread popular interest in artificial intelligence, including across the humanitarian community. Luke traces advances in large language models and shares his views on where we are today and where the future of generative AI and the uses of large language models are headed. Luke also touches on top-down pressure on engineering teams to leverage AI, HelixML’s involvement in helping companies deploy large language models locally on their own infrastructure, and Anthropic’s Model Context Protocol (MCP) which standardizes how AI models connect with different data sources and tools, as well as the future of MLOps.
    Exibir mais Exibir menos
    21 minutos
  • The New Commons Challenge: Advancing AI for Public Good through Data Commons

    The New Commons Challenge: Advancing AI for Public Good through Data Commons
    May 2 2025
    Stefaan Verhulst and Andrew Schroeder speak with Brent Phillips, Humanitarian AI Today podcast producer, about the New Commons Challenge (https://newcommons.ai/) focusing advancing AI for public good through data commons. Dr. Stefaan G. Verhulst is an expert in using data and technology for social impact. He is the Co-Founder of several research organizations including the Governance Laboratory (GovLab) at New York University and The DataTank based in Brussels. Dr. Andrew Schroeder is the Vice President of Research and Analysis for Direct Relief and the Co-Director of CrisisReady. The Open Data Policy Lab, a collaboration between The GovLab and Microsoft, launched the New Commons Challenge initiative to advance the responsible re-use of data for AI-driven solutions that enhance local decision-making and humanitarian response. The Challenge will award two winning institutions $100,000 each to develop data commons that fuel responsible AI innovation in these critical areas. The Challenge builds on the Open Data Policy Lab’s recent report, “Blueprint to Unlock New Data Commons for AI,” which advocates for creating collaboratively governed data ecosystems that support responsible AI development. The Challenge is backed by leading institutions committed to ethical AI and open data. Partners include Direct Relief / CrisisReady, and Harvard Law School Library, and UNESCO as international observer, which will provide subject-matter expertise and evaluation support. Stefaan, Andrew and Brent discuss the New Commons Challenges in detail and touch on data commons for AI, disaster relief and local decision making, AI governance, data sharing architectures and data re-use, advances in artificial intelligence and human-AI interaction, and the intersection of collective intelligence and artificial intelligence.
    Exibir mais Exibir menos
    39 minutos