Foundation/Mozilla Education/Events
Jump to navigation
Jump to search
2026 Events and Convenings
August
| When Research Meets Reality: AI Engineering in Africa (Lagos, Nigeria; 5–7 August 2026; 100+ Participants) In collaboration with DataSpires, we convened a deep-dive discussion with over 100 researchers and startup founders at the Deep Learning Indaba 2026, held at Pan-Atlantic University under the conference theme of "Sovereign Intelligence," to explore the challenges and opportunities in translating African AI research into viable products and innovations. Rather than dwelling on technical requirements alone, the practitioner-led conversation examined the broader ecosystem conditions—commercialisation pathways, skills development, funding models, governance frameworks, and market readiness—that determine whether promising research moves beyond experimentation into real-world deployment and impact. Through short lightning talks from researchers, founders, and product leaders, a moderated panel, and interactive audience dialogue, participants surfaced where opportunities exist, where bottlenecks emerge, and what support structures African AI innovation pathways require—including how infrastructure constraints shape research ambitions and lessons from building and scaling AI solutions in African contexts. Running alongside the session, a revised EQUAL Compute community survey—building on the 2025 effort with DataSpires at the Indaba in Rwanda—mapped computing usage and needs across the wider cohort to better understand the relationship between research ambitions, infrastructure constraints, and innovation outcomes. Key Takeaway: The journey from research to real-world application is many times more complex than compute access alone—infrastructure, skills, funding, and institutional support interact in ways that make ecosystem conditions, not just technical capacity, the decisive factor in whether African AI research becomes deployable products and public value. |
July
| Human Scale AI Buildathon Launch (Bangalore, India; 31 July 2026; 100 Participants) We formally inaugurated the Human Scale AI Buildathon at CHRIST University's Bannerghatta Road Campus in Bangalore, launching a six-month national student-builder journey focused on Human Scale AI for Indian contexts—local languages, low-resource settings, limited connectivity, and communities often left out of technology design. The ceremony opened with a fusion performance, followed by the ceremonial lamp lighting; Dr. Anil Joseph Pinto, Registrar, spoke on technology, society, and ethics, and Dr. Fr. Jossy P. George, Director of the BGR Campus, felicitated the guests and set out the institutional vision. In the Mozilla keynote we made the case for Human Scale AI, for young people as builders, and for building AI that expands human capability rather than creating dependency—a theme carried into a plenary on "Technology, Society and Ethics in the AI Era", and an afternoon Education Assembly workshop, "Capability or Dependency?", on AI usage, dependency, and user autonomy. Key Takeaway: Growing and continually investing with existing community partners over time, rather than starting fresh each time, is what allows implementation to scale. | |
| Responsible Computing Beyond the Classroom Paper Presentation (Cape Town, South Africa; 13–15 July 2026; # Participants) Dr. Chao Mbogho presented a peer-reviewed paper — "Responsible Computing Beyond the Classroom: Student, Faculty, and Institutional Outcomes Across Eight Kenyan Universities" (Mbogho, Azeka & Edward-Gill) — at the South African Institute of Computer Scientists and Information Technologists conference, held at the University of Cape Town Graduate School of Business. Using the Kirkpatrick evaluation model and drawing on institutional reports, a faculty survey, and student testimonies, the paper looks past what students learn inside a responsible-computing (RC) course to what RC education changes afterwards, reporting on a two-year rollout of Mozilla's Responsible Computing Challenge that reached 2,847 students across eight Kenyan universities. It documents RC work spilling well beyond the classroom: student projects becoming deployed products and ventures, faculty earning publications, promotions, and policy roles, and institutions adopting new RC-aligned policies and courses — gains that depended on faculty champions with leadership buy-in. Key Takeaway: Sustained RC integration drives systemic change across students, faculty, and institutions that in-classroom evaluation misses, and the paper offers a transferable three-level framework for measuring it — evidence still rare from African computing-education research. | |
| Why Open Badges Still Matter (Boulder, US; 13–15 July 2026, 100 Participants) We joined the digital-credentialing community at CU Boulder for the 2026 Badge Summit and supported the opening "Open Badges Then and Now" pre-conference with the Learning Economy Foundation — a fitting thread, since the Open Badges standard the field now builds on began at Mozilla. Across three days the conversation centered on making credentials trustworthy and learner-owned rather than merely plentiful: centering learner voice and agency in ecosystem design, so that a good badge helps a learner tell their story; open-source infrastructure for the full credential lifecycle — award, store, share, verify — including learner wallets and learner-controlled credential data through the Digital Credential Commons; and democratizing durable skills through open competency libraries with performance-based, rubric-scored assessment (Education Design Lab on Open edX). The sharpest provocation came from the "Good Badge, Bad Badge" session: if a learner cannot name the skill, explain what they did to earn it, and say why it matters to an employer, the badge has failed them — so the real design question is not what metadata to attach but what claim we are making about a learner, and what evidence makes that claim trustworthy. Key Takeaway: The field's frontier is no longer issuing more badges but earning trust — designing credentials around verifiable, learner-owned claims and strong signals, so that "not just what you did, but what you can do" becomes legible to employers and learners alike. | |
| Democracy & AI Hackathon (Nairobi, Kenya; 2–4 July 2026; 32 Participants) We ran a three-day hackathon with KamiLimu at iHub in Nairobi, bringing together 32 university students in 16 teams from 10 universities to build AI tools that strengthen democracy—across government accountability, civic literacy, and information integrity. After judges' training and a day of pre-hackathon preparation, teams moved through a plenary requirements session, six hours of building, and pitches to a four-judge panel, with every student who started the day present to pitch at the end. Through our involvement, we kept responsible computing central to the rubric and the mentoring, even as it proved the hardest segment for every team. Key Takeaway: This kind of applied, well-supported build genuinely shifts students from consumers of AI into responsible builders of it: only 2 of 31 had ever built AI for a democratic problem before, yet 90% left more confident building responsible, democracy-focused AI, 29 of 30 rated the experience better than other hackathons they had attended, and seven teams advanced into six weeks of mentored building toward the 21 August finals. |
June
May
| Creative Computing in Education Showcase (New York, US; 18 May 2026; 17 Participants) We help lead a semester-long creative computing course partnered with the College of Staten Island for a cohort of K–12 teachers on sabbatical, learning from them how responsible programming, data visualization, and interdisciplinary computing could come alive across ELA, the arts, and social studies. The course culminated in a multi-week hackathon in which teams asked what the classroom of 2046 should look like, ideated solutions, and built capstone projects—producing zines and classroom-ready lesson plans to carry back to their own schools, including a student-media unit interrogating whether AI companions empower adolescent agency or quietly ghostwrite it. Through our involvement, we kept responsible computing and ethics at the center, encouraging teachers to examine AI's role in assessment, student agency, and wellbeing rather than adopt it uncritically. Key Takeaway: Educators consistently imagined AI as an augmenting co-teacher rather than a replacement—with the human teacher remaining the lead, the emotional anchor, and the final decision-maker in the classroom. | |
| Africa Compute Initiative Inception Workshop (Cape Town, South Africa; 7–8 May 2026; 35 Participants) We co-organised a two-day inception workshop with the University of Cape Town to launch the Africa Compute Initiative (ACI)—a compute project under the EQUAL Compute Network and the wider AI4D program, with funding from the Government of Canada and FCDO and additional support from the Gates Foundation and other partners. 35 African AI experts came together around a central question: what would it take to build compute infrastructure that genuinely serves African AI research and innovation? Day One examined the supply side—technical architecture, governance, and allocation trade-offs—while Day Two turned to demand, surfacing a deep well of suppressed research across malaria drug discovery, African-language NLP, sign-language translation, climate forecasting, and mother-tongue court transcription that researchers simply aren't pursuing because they assume compute is out of reach. Case studies from UCT and Rwanda's HPC systems, Sunbird AI's Sunflower models, and DataSpires' idle-GPU marketplace grounded the discussion in practice. Key Takeaway: GPU access alone is not enough—compute must be paired with storage, tooling, training, and transparent governance—and that national data-sovereignty laws make a single Cape Town hub structurally insufficient, pointing ACI toward a distributed, sovereignty-compliant model. Article: African Compute Initiative's inception workshop advances shared vision for AI infrastructure in Africa | |
| Democracy & AI Student Virtual Launch (Online, Kenya; 6 May 2026; 86 Participants) We opened the Democracy & AI Student Builder Program — a multi-phase buildathon run with KamiLimu and Responsible Computing faculty across Kenyan universities — with an online Assembly that treated theme-setting as construction rather than broadcast. Across 13 breakout rooms, 86 students surfaced the democratic challenges they see in Kenya, and we clustered those patterns into three student-shaped sub-themes — civic literacy, government accountability, and information integrity. The Assembly fed directly into a problem-ideation phase in which 68 students in 34 teams from 17 universities (10 of them rural or peri-urban, with a near-even gender split) researched, created problem statements using a scaffolded template, before ten trained reviewers began returning detailed feedback to every team. Key Takeaway: Co-creating the themes with students rather than handing them down produced unusually deep buy-in — zero detractors and a +81 NPS — and a well-scaffolded template with worked examples proved the great equaliser, letting students from seventeen very different universities produce comparably rigorous work. | |
| Reframing Education as Infrastructure (New York, US; 1 May 2026; ~300 Participants) At the Transforming Global Education Summit in the UN, we moderated a panel with Diana Krogh (The LEGO Group), Melanie Dukes (LearnerStudio), and Simone Ravaioli (Instructure) that reframed education from a service to be delivered into national digital infrastructure that nations own and govern — arguing that the service framing stifles innovation and, by over-indexing on formal recognition, excludes. Speakers mapped what a sovereign operating layer requires (connecting formal and informal learning, feedback beyond math and literacy scores, real-world signalling) and grounded it in practice through LEGO's Build the Change, open-source digital well-being tools like RITIC, and MyEquals, the decentralised credentialing network scaling across Australia, New Zealand, and the Pacific Islands. Key Takeaway: Large-scale learning ecosystems are first a coordination problem at the constitutional layer, prioritizing recognition over instruction delivery, measuring curiosity and agency rather than test scores, and engaging children and youth directly as decision-makers. |
April
| Mozilla Education K-12 Student Showcase (Online; 24 April 2026; 24 Participants) Students presented the results of a semester long project that asked them to survey their peers, research a problem, and design either a tech tool or a tech-focused education program to support young people as they move into college and the workforce. The eight teams tackled problems drawn largely from their own lives: an AI triage assistant (Nebula) and a subscription telehealth app (Health Square) to cut healthcare wait times; SafeVoice, an anonymous cyberbullying-reporting app that routes urgent cases to counsellors; EduBalance, a classroom platform pairing focus tools with project-based learning; Readify, an app matching teens to books at their reading level to rebuild literacy; a digital-detox proposal addressing phone addiction; a Shark Tank-style pitch to close NYC's teen-employment gap; and a student-survey analysis of bias and fairness in school discipline. Key Takeaway: Given the framing of responsible technology design and asked to solve problems in their own communities, students gravitated to the issues professional product teams often sideline — equitable access, privacy, mental health, and literacy — and built those values in from the start rather than bolting them on. | |
| The Future of AI in Higher Education (Los Angeles, US; 9 April 2026; 8 Participants) We hosted a focus group of education-research experts at the American Educational Research Association's annual meeting, convening faculty and doctoral researchers to examine how AI is reshaping higher education. The conversation moved from curriculum design to community, disciplinary boundaries, early education, and ethics—returning again and again to the idea that AI works best as a temporary scaffold, one students use to build internal knowledge before the support fades and they connect ideas on their own, and that this has to be designed into the curriculum intentionally rather than retrofitted onto it. Participants also pushed on how AI is redrawing who gets access to which fields—questioning whether calculus should remain a gatekeeper to computer science, and how the humanities and arts shift as critical thinking becomes the durable core skill. Key Takeaway: Community is not a soft add-on but the load-bearing infrastructure for learning with AI—investment should flow into building trusted cohorts, not only tools, since one-person, one-session AI use works against exactly the kind of connection that makes education durable. |
March
| The Impact of Artificial Intelligence on Education and Workforce Trajectories in Tech (Washington D.C., US; 30-31 March 2026; ~1,200 Participants) We helped plan a two-day workshop on The Impact of Artificial Intelligence on Education and Workforce Trajectories in Tech at the National Academies of Sciences, which brought together over 1,200 participants from academia and industry for a national conversation centered on how we can better prepare students to succeed in an AI-driven tech workforce. Through our involvement, we ensured that ethics and responsible AI remained central to the discussion, including placing several of our awardees on panels — and our early investment in Harvard's Embedded EthiCS program generated significant attention. Key Takeaway: Students shouldn't just consume AI tools — they need to move from users to makers, building durable judgment, ethical reasoning, and critical-thinking "muscle" that AI cannot supplant. | |
| Future of Personal AI (Seattle, US; 11-12 March 2026; 100 Participants) We co-hosted a Builder-a-thon with the Gates Foundation, Stanford HAI, AWS, and Slalom to bring together researchers, universities, industry, and nonprofit organizations around the future of personal AI in education. Builder teams developed interoperable, person-centered prototypes—including portable memory systems, learner profiles, and privacy-preserving frameworks—that prioritized user agency, trust, consent, and interoperability. Participants reached broad consensus on the need for open standards, user-governed personal memory, longitudinal evaluation methods, and governance frameworks that ensure AI systems remain transparent, portable, and accountable across education and other public-interest domains. Key Takeaway: Students and researchers can make meaningful contributions to the development of next-generation AI infrastructure alongside industry leaders. |
February
| AI, Learning, and the Future of Academic Frameworks (Bengaluru, India; 7 February 2026; 13 Participants) We facilitated a roundtable dialogue at CHRIST (Deemed to be University), co-hosted with the university's Department of English and Cultural Studies, convening students, PhD scholars, faculty, and industry professionals to examine how AI is reshaping curriculum, assessment, and human capability. The conversation moved through four questions—whether AI accelerates cognitive decline, whether it can rival faculty as a learning facilitator, how to ensure equitable access in under-resourced settings, and what its rise means for employability. Across four questions—AI's effect on cognition, its role versus faculty, its accessibility, and its impact on employability—participants repeatedly concluded that AI itself is fairly neutral, with outcomes hinging on how ethically and deliberately it is integrated into research, teaching, and policy. Key Takeaway: AI can democratize knowledge but cannot replace the human interaction and motivation at the heart of education, while access gaps and freemium models risk deepening existing inequalities. |
January
| The Final CS K-12 Standards Revision Meeting (San Diego, US; 20 January 2026; 27 Participants) Steven Azeka served on the Computer Science Teachers Association (CSTA) Standards Writing Committee representing the Mozilla Foundation alongside a group of K–12 CS leaders, through the multi-year revision of the foundational K–12 computer science standards for the United States — standards expected to be adopted in some form by nearly all fifty states. With the final round of edits now complete, our contribution focused on embedding responsible computing principles for the age of AI and elevating interdisciplinary, humanistic approaches to the discipline. Through our involvement, we worked to ensure responsible computing is treated not as an add-on but as a foundational expectation for every K–12 student in the country. Key Takeaway: Shaping the standards themselves — the layer nearly every state builds its CS curriculum on — is among the highest-leverage ways to make responsible, human-centred computing the default for a generation of U.S. students rather than optional enrichment. Standards Link: A Vision for PK–12 Computer Science Education |
2025 Events and Convenings
November
| Roundtable on AI Framework Development for Academia (Kerala, India; 22 November 2025; 60 Participants) Our Responsible Computing Challenge team in India, with Tinkerhub and IIIT Kottayam, convened students, faculty, curriculum designers, and trainers to examine how AI is reshaping learning, assessment, and employment across Indian higher education. Participants welcomed AI's power to democratize knowledge and speed up learning, but voiced sharp concern that heavy reliance is eroding critical thinking, attention, and memory, and that memory-heavy end-semester exams have fallen badly out of step with how students now actually work — one student describing a "foggy" mind, one faculty member calling the gap between AI-skilled students and decades-old evaluation systems unsustainable. The discussion surfaced a widening divide between digital-native students and their teachers, deep student anxiety about automation and placements, and interlocking infrastructure, access, skill, and pedagogy gaps — alongside real opportunities in AI tutors, simulation labs, flipped classrooms, and continuous, judgment-based assessment. Key Takeaway: AI is improving faster than institutions can adapt, and the response should be structural rather than restrictive — treat AI as a facilitator to design around, not to ban: rebuild assessment around judgment and application (including AI-permitted open-book exams), embed AI literacy early, invest heavily in faculty skilling, and widen equitable access so that AI narrows rather than deepens existing divides. |
October
| Cultivating Africa's AI Talent, Skills, Education and Cultural Innovation Panel (Cape Town, South Africa; 1 October 2025; 30 Participants) At the G20 AI for Africa Conference — co-hosted by the African Union, the South African G20 Presidency, and UNESCO — Steven Azeka, joined a UNESCO-moderated panel on how Africa can build an inclusive AI ecosystem and turn the "brain drain" into a "brain gain." Alongside representatives from Namibia's National Commission on Research, Science & Technology, Kenya's ICT ministry and UNESCO's SPARK Alliance, the Smart Africa Secretariat, KNUST's Responsible AI Lab in Ghana, and the Congo AI Research Centre, Mozilla made the case that open-source AI is a strategic advantage for the continent: open models can be fine-tuned to local languages and needs in ways that closed frontier models built in the Global North will not, provided their development is approached critically and responsibly. Drawing on the Responsible Computing Challenge — which invests in university faculty to build responsible-computing curricula, already running in Kenya and expanding to South Africa — the panel argued for a "full-stack" view of the AI workforce that reaches beyond data scientists and engineers to the wider ecosystem of people who steward data, build sustainable compute infrastructure, and train others, echoing the call for "home grown solutions addressing home grown problems." Key Takeaway: Africa can convert brain drain into brain gain by building an ethical, full-stack AI ecosystem rather than importing one — pairing open-source models fine-tuned to local needs with responsible-computing skills taught in universities before graduates enter the workforce — so that a homegrown AI economy includes and rewards far more than engineers and content moderators. |
May
| EQUAL Compute Network Implementation Workshop "Impshop" (Ottawa, Canada; 22–23 May 2025; 16 Participants) We joined funders and core partners convened by the International Development Research Centre (IDRC) in Ottawa for a two-day implementation workshop — the third in the EQUAL Compute Network series, following Paris and Kigali — to turn the Network's vision into a working investment strategy and operational model. Drawing on demand modelling that showed cloud compute can cost an African researcher many times what it costs a peer in Europe, participants converged on a "menu of options" approach: widening compute access across the global South starting with Africa, prioritising scalable use cases tied to the Sustainable Development Goals, and supporting dedicated country-level clusters where they make sense. The group sketched a structure of regional hubs — coordinated by a funders' collaborative with an independent advisory council — responsible for demand aggregation, procurement and matchmaking, capacity building, policy advocacy, and monitoring and learning, and agreed that access should be tiered by user type, with full subsidy for public-interest research, partial subsidy for NGOs and startups, and cost-recovery for commercial use. Participants began identifying near-term "lighthouse" pilots, including an under-utilised GPU cluster at the University of Cape Town, and aligned on positioning the initiative ahead of the G7 and G20. Key Takeaway: The initiative should advance not as a single infrastructure bet but as a demand-led "menu" of compute options — foreign cloud and HPC alongside small and regional local clusters — delivered through regional hubs and proven out on a handful of concrete lighthouse use cases; success is measured socially, in inclusion, sustainability, and equitable access, and long-term viability depends on a tiered, blended funding model that weans the Network off short-term donor cycles. |
April
| EQUAL Compute Network Workshop II (Kigali, Rwanda; 1 April 2025; 29 Participants) At the Global AI Summit on Africa, we reconvened the EQUAL Compute Network together with the International Development Research Centre (IDRC) and the Alliance for AI, to move from shared vision toward a concrete model and structure for the Network. Building on the Paris strategic priorities of equitable access, upskilling, and compute governance, participants worked through breakout sessions that pressure-tested a proposed Network structure and its governance — who can access compute resources and how, what roles different actors should play, and how the model can sustain supply over time — before turning to a top-level implementation plan and timelines on the road to a final implementation report. Key Takeaway: The Network's design should be demand-led rather than technology- or funder-led — scanning for real, on-the-ground needs and guarding against the bias of letting funders or vendors define them — while committing to equitable, tiered access that reaches underrepresented countries, and to a long-term sustainability model that reduces donor dependency by working through institutions already in place, such as universities and energy providers, and prioritizing scalable, African-specific use cases. |
February
| EQUAL Compute Network Workshop I (Paris, France; 12 February 2025; 40 Participants) On the margins of the French AI Action Summit, we convened the inaugural EQUAL Compute Network Workshop together with the International Development Research Centre (IDRC) and the Alliance for AI, bringing funders, African universities and AI labs, industry compute providers, civil society, and government representatives into one room to advance the design of a network that would narrow the AI compute divide between the global South and global North. After context-setting on the state of Africa's compute infrastructure and skills, participants worked through breakout sessions to pressure-test a notional network structure — how it should be governed, who should benefit, and how to aggregate demand across cloud and on-prem — and to define which constituents should be involved and what roles they should play, all anchored in principles of being inclusive, responsible, and green. Key Takeaway: The Network should be built as independent regional hubs rather than a single central facility, coordinated by an independent operating entity backed by an advisory council and a program-management team; to endure, it must be financially self-sustaining through a mix of funding sources — donors, venture capital, government, and public-private partnerships — and it must stay anchored in locally relevant use cases, data sovereignty, and equitable resource distribution, engaging major cloud providers while building local compute capacity. |
