Do Education Ministries Need Skunkworks to Get Ahead of AI?

This is part of a series hosted by the AI Observatory, offering perspectives on key issues in Education in the Age of AI. We hope you find this opinion piece thought-provoking and encourage further discussion on the topic.

Education ministries around the world are facing a tough balancing act: delivering more, faster—with less people than they need, tighter budgets, and rising expectations. All while keeping pace with the most significant wave of technological change in a generation.

AI is here and it’s changing how we work, learn, and make decisions. And yet most public delivery systems, particularly in education, aren’t set up to respond to this kind of change, let alone shape it.

One thing is clear: if we were designing ministries from scratch today, they’d look nothing like they do now. They’d be leaner, tech-enabled, and – in the age of AI – increasingly automated. But the reality is more complex. Most ministries are contending with deeply embedded legacy systems and layers of incumbency that span the entire education delivery chain. It’s a lot to untangle. We’ve been reflecting on how ministries can best respond – whether by catching up or getting ahead – and what it really takes to ‘turn the tanker’.

So here’s a question we’ve been exploring: What if every Ministry of Education had its own Skunkworks?

What’s a Skunkworks, anyway?

A Skunkworks is a small, semi-autonomous team given some degree of freedom within a bureaucracy in order to work quickly on a forward-looking initiative outside of normal day-to-day operations, with a direct line of report to the chief executive

The term comes from Lockheed Martin’s famous Skunkworks division. In 1943, during WWII, the U.S. Army Air Forces asked Lockheed to build a jet fighter in just 150 days. The job went to a small team, working outside the usual bureaucracy. They delivered the XP-80 Shooting Star in 143 days.

Since then, the Skunkworks model—small, protected teams given the freedom to work fast and differently—has cropped up across government around the world:

• 🇬🇧 UK: the Government Digital Service (GDS) started as a small team inside the Cabinet Office. They rebuilt www.GOV.UK using agile methods, showing that government can deliver modern, user-centred services if the environment allows it. In its early years it was seen as a paradigm shifting initiative. It’s currently going through another era of transformation, combining central digital and data teams, AI incubators, geospatial teams, and responsible tech adoption.
• 🇪🇹Ethiopia: the Artificial Intelligence Institute is a national institute exploring various domains of AI and its applications across government from health, agriculture and finance.
• 🇮🇩Indonesia: INA DIGITAL is grounded in the ambition of GovTech transformation and delivers digital public services, including for education.
• 🇵🇭Philippines: AI Center for Education (E-CAIR) is the first-ever hub for AI research in education in the Philippines and is focused on revitalising basic education by developing AI-driven tools that enhance teaching, learning, and school administration.
• 🇺🇾Uruguay:Ceibal is the national center of educational innovation and digital technologies and promotes technology integration to improve the quality of learning.
• 🇵🇪Peru: MineduLAB is an innovation lab for education policy and the first government innovation unit in Latin America to use Randomised Controlled Trials as its main strategy for policy innovation (Innovations for Policy Action case study). The Lab was “able to produce timely and relevant results at very low cost, which in turn allowed securing the support of high-level officials within the ministry”. (Hernández-Agramonte & Iglesias, 2023)
• 🇬🇧 UK: UK NHS AI Skunkworks looked for new ways to use AI for driving forward the early adoption of technology to support health, in both clinical and business contexts. The team provided free short-term expertise and resources to public sector health and social care organisations to support AI projects and develop capability.
• In the private sector, Skunkworks have been set up in Google, Apple and others. At Apple, for example, the Skunkworks codenamed Project Purple was responsible for the iPhone.

These examples show us what’s possible when large institutions carve out space for experimentation, but also the different paths that can lead there. GDS focused on exploring the broader opportunity of digital transformation across government, securing top-down support for “revolution, not evolution” from the Cabinet Office.

What would be the point of an education ministry Skunkworks?

Today, inside most ministries we work with, information moves slowly. Feedback loops are weak. Frontline staff are often out of the loop or overwhelmed. In a world of increasing automation and predication, these systems are still stuck in manual mode.

A Skunkworks approach would mean not having to don’t have to wait for the perfect reform moment. Ministries can take parallel strategic paths to create their own Skunkworks by setting up a small, internal team with permission to work differently, test new tools, and build momentum from the inside out.

This kind of approach could help us to go beyond chatbots or dashboards and rethink how ministries make decisions, design systems, allocate resources, and adapt to live data, updating the delivery engine itself.

Think of it like a small in-house team with 3–6 people, protected from bureaucracy, focused on solving real delivery challenges using modern tools and methods. In fact, here are some of the questions we are getting asked today:

• Can predictive AI help officials in Kenya make smarter secondary school student allocations and address climate-related school disruptions?
• Could generative AI boost the content and engagement of Nigeria’s learning management system platform?
• How might we rethink procurement practices to ensure choices are evidence-based on allow flexibility and implementation research as technologies scale?

The team would start by exploring these questions, but they’d move quickly from theory to build things, test them, learn fast, and share results. All while staying connected to the Ministry’s core priorities.

Creating skunkworks in education ministries means we could test whether AI-enabled tools help improve the speed and quality of delivery decisions? Ministries are making critical choices about resource allocation, planning, and implementation with limited, often delayed data.

For the sector, Skunkworks could help us to generate empirical evidence around a widely held assumption that AI can help government teams make better decisions, faster.

Further reading:

• What do AI futures mean for human connection?
• How to unlock learning outcomes in the age of AI?
• Test time: Humanity’s last exam and education in AI
• What we’re reading and hearing about AI in education
  

---

These reflections are some of the important ones we are thinking about in the AI observatory, and we invite you all to consider them alongside us. We’ll be sharing more views and opinions from experts in the space, and we want to hear your thoughts as well! Join the conversation on LinkedIn.