One million people. Not developers. Not data scientists. Not AI engineers — people. The UAE and Lebanon have just launched an initiative to equip one million Lebanese people with practical skills in generative AI and prompt engineering, and I think the most important part of this announcement isn't the number. It's what the initiative represents.
Because we are beginning to see a fundamental shift in how countries think about national capability. For decades, governments invested in roads, airports, ports, schools, universities, telecommunications, and digital infrastructure. Today another layer is being added: AI capability. And the UAE is increasingly treating AI literacy not simply as a technology skill — but as infrastructure for an entire society.
What was announced
On the sidelines of the 81st session of the United Nations General Assembly in New York, H.H. Sheikh Abdullah bin Zayed Al Nahyan, UAE Deputy Prime Minister and Minister of Foreign Affairs, met Lebanese Prime Minister Dr. Nawaf Salam. During the meeting, the UAE and Lebanon announced the One Million Lebanese AI Experts initiative, operating under the UAE's Government Experience Exchange Programme in cooperation with the Government of Lebanon.
Its goal is to equip one million Lebanese people — across different ages, backgrounds and roles, not merely technology professionals — with fundamental and practical skills in generative AI and prompt engineering: using AI to accomplish tasks, increase productivity, improve working methods, develop ideas, solve problems, and support better government services. That distinction matters. This is not about creating one million AI engineers. It is about creating a society capable of working with AI.
AI literacy is becoming the new digital literacy
Think about what happened with computers. Thirty years ago, knowing how to use a computer was a specialist skill. Then it became an office skill, then a basic professional skill, and eventually a basic life skill. The same happened with the internet, then smartphones. I believe we are watching the same transition happen much faster with AI. Today we still say "AI skills" — soon that phrase may sound strange, because using AI may become as fundamental to work as searching Google, writing an email, using Excel, or joining a video call. The question will no longer be "do you know AI?" It will become how effectively can you work with AI?
This changes the definition of workforce productivity
Imagine two employees — same education, same experience, same role, same working hours. One works conventionally; the other knows how to research, analyse documents, draft, summarise, generate presentations, automate repetitive work, build simple agents, translate, and prepare decisions with AI. The second employee may not simply be 20% more productive — in some tasks, the difference could be multiples. That means AI literacy can eventually become an economic productivity issue, not merely a technology issue.
The UAE has already been building this model at home
This announcement isn't happening in isolation. The UAE has already been pursuing large-scale AI capability development domestically, targeting AI skills across government employees and the wider workforce, while simultaneously moving significant parts of government toward Agentic AI, where AI systems increasingly execute workflows and support decisions rather than merely answer questions. So there are two transformations happening at once: making the machines more capable, and making the people more capable of working with the machines. You need both — deploying advanced AI into a workforce that doesn't understand how to use it produces limited value, and training people without redesigning systems produces limited value too. Real transformation happens when people, process, data and AI change together.
Here is what I would do next: five levels of AI capability
Teaching one million people prompt engineering is an excellent starting point, but I wouldn't stop there — because the AI skill required in 2027 will be very different from the skill required in 2024. The first generation was "prompt AI" (ask the machine); the next is "work with AI," then "delegate to AI," then "manage AI," and eventually "supervise digital workers." I would structure national AI capability in levels:
Level 1 — AI Citizen
Every person should understand what generative AI is, how to prompt, how to verify outputs, how to protect personal information, how to recognise hallucinations, and how to use AI responsibly. This is basic AI literacy.
Level 2 — AI Professional
Apply AI to real work — finance, marketing, education, healthcare, engineering, tourism, retail, government, legal, media, entrepreneurship. The objective isn't "learn ChatGPT." The objective is to redesign how you do your job.
Level 3 — AI Builder
Now people learn to create — agents, automations, AI workflows, knowledge systems, AI-powered applications, APIs, low-code solutions, business prototypes. AI literacy becomes AI creation.
Level 4 — AI Manager
This is where I think the next enormous skills gap will emerge. Soon managers may supervise humans, AI agents, and hybrid teams — understanding agent performance, permissions, quality, human escalation, AI economics, risk, governance, and when (and when not) to automate. The manager of 2030 may not only manage 10 people; they may manage 10 people plus 200 AI agents. That requires a completely different management skill set.
Level 5 — AI Entrepreneur
This is where a national programme can become economically transformative — don't only teach people to consume AI, teach them to build businesses with AI. If just 1% of one million learners eventually launch an AI-enabled business, that's 10,000 entrepreneurs. If another 5% redesign workflows inside existing companies, that's 50,000 transformation champions. If thousands build local Arabic AI tools, education products, tourism platforms, government solutions, healthcare applications, fintech, or SME automation, the programme stops being a training initiative — it becomes economic infrastructure.
Lebanon has something particularly valuable
Lebanon has a large pool of entrepreneurs, engineers, developers, designers, educators, professionals, students, and a globally distributed diaspora. AI dramatically reduces the amount of capital required to create certain kinds of digital businesses — a founder no longer necessarily needs 20 developers, a large office, huge infrastructure, and millions in early capital. A small team equipped with AI can increasingly research, design, code, market, analyse, sell, support customers, create content, and operate internationally. That changes the economics of entrepreneurship. Imagine connecting Lebanese entrepreneurial talent with UAE AI experience and global digital markets — that could become much bigger than training.
Government should measure outcomes, not certificates
This is where I would be very disciplined. Don't measure success only as "one million people trained" — that is an activity metric. Measure what changed: how many people use AI weekly? How many working hours were saved? How many companies adopted AI? How many AI businesses were created? How many new jobs emerged? How many AI products were exported? That turns AI training into AI economic impact.
I would give every participant one real problem after completing the programme — not just a certificate. Improve one process at work. Build one AI workflow. Save five hours per week. Prototype one business. Teach five other people. Then measure the results. One million people × one AI improvement is potentially one million experiments in productivity.
This is also a new form of government cooperation
Historically, countries exported capital, infrastructure, technical expertise, development assistance, and institutional knowledge. Today we are beginning to see another export: AI capability. The UAE's Government Experience Exchange Programme is already built around transferring government knowledge and institutional experience to partner countries — now AI capability becomes part of that exchange. This creates an interesting model: AI diplomacy through capability building. Not simply "here is technology," but "here is the capability to use technology." Technology becomes obsolete; human capability compounds.
Imagine where this model could go — One Million Lebanese AI Experts, then perhaps similar capability programmes elsewhere, tailored to what each country needs: AI for government, agriculture, education, healthcare, SMEs, tourism, or entrepreneurship. The UAE could increasingly export something it has spent years developing: the operating model for AI-enabled government and economies. That may ultimately be more valuable than exporting a single technology platform.
The bigger government lesson
Governments around the world are racing to buy GPUs, models, cloud infrastructure, data centres, and AI platforms. But perhaps one of the highest-return investments is much simpler: teach the population how to use the intelligence that already exists. Infrastructure without capability doesn't automatically produce productivity — a country can have access to the world's most powerful AI models, but if employees don't know how to use them, managers don't redesign workflows, and companies don't change operating models, the economic impact remains limited. The AI race isn't only "who has the best model?" It is increasingly who can turn AI into the most human capability?
My question is not whether one million people can learn AI — they can. My question is what happens after they learn it. Can one million learners become 100,000 advanced users, 50,000 transformation leaders, 10,000 builders, 5,000 entrepreneurs, 1,000 new AI companies? Because the ultimate KPI should not be one million certificates — it should be one million people capable of doing something they could not do before.
Countries once built national competitiveness by teaching millions of people to read, write, use computers, and use the internet. The next chapter may be teaching entire populations to think with AI, work with AI, build with AI, manage AI, and eventually create economic value with AI. That's when AI stops being a technology programme. It becomes national capability — and that is a much bigger ambition.