For decades, we told people that if they wanted to succeed in a global economy, they needed to speak English. You could be an engineer in Seoul, an entrepreneur in Abu Dhabi, a civil servant in Singapore or a banker in Paris; English might not be your first language, but it must be your second if you want to compete and trade globally as it became, over the past century, the common language of international business, technology, academia and diplomacy.
Today, we are entering a similar moment, except the new second language is not English. It is AI.
This may sound exaggerated until we consider how quickly the workplace is changing. The World Economic Forum estimates that 86 per cent of employers expect AI and information-processing technologies to transform their businesses by 2030, while 39 per cent of workers’ existing skills will change or become outdated. If the global workforce consisted of 100 people, 59 would need training before the end of the decade, with 11 unlikely to receive it. AI and big data already rank as the fastest-growing skills employers expect to need.
Singapore sees this challenge, knows what this means and is moving to address this coming wave. Last month, its Skills and Workforce Development Agency launched a catalogue of more than 200 AI courses. More interestingly, eligible Singaporean citizens aged 18 and above who complete selected courses can receive six months of complimentary access to a premium AI tool, including ChatGPT Plus or Google AI Pro. Prime Minister Lawrence Wong described the initiative as part of a broader attempt to prepare workers for jobs whose skills and requirements are changing rapidly. Singapore wants AI to be its entire populations’ second language.
The second part matters more than it first appears, because Singapore is not simply telling citizens to learn about AI; it is giving them the opportunity to live with it. Take a course, get access to the technology, use it every day, make mistakes and discover where it helps and where it does not. Nobody becomes fluent in French by attending a two-hour presentation about French, because languages are learnt through repeated use until translating in your head is no longer necessary. AI will be similar.
Most people today are probably somewhere around tourist-level AI. They can ask ChatGPT a question, rewrite an email, summarise a document or generate an image, which is useful but roughly equivalent to knowing how to order coffee and ask for directions. AI fluency is different because it means instinctively asking, before beginning a task, whether AI could help you do it better.
Before researching an issue, you use AI to map the field; before a meeting, you ask it to challenge your assumptions; before writing a report, you interrogate the evidence. You use it to analyse data, compare options, simulate scenarios, generate alternatives and critique your own thinking. Eventually, AI becomes less like an application you occasionally open and more like an additional layer of cognition sitting alongside you.
There is already evidence that this matters. In a field experiment involving 7,137 knowledge workers across 66 companies, employees given an AI assistant integrated into their everyday work spent about two fewer hours each week on email. Meanwhile, LinkedIn reported that the number of AI-literacy skills added by its members had risen 177 per cent since 2023, while more than half of hiring managers surveyed said they would not hire someone without AI literacy.
Singapore’s approach goes further by trying to measure fluency. Its AI Ready Quiz, developed with the Singapore Institute of Technology, takes about 15 to 20 minutes and places users along a progression from AI Aware to AI Literate and AI Fluent, before recommending appropriate learning. The equivalent in the analogue world was English language tests like Toefl or IELTS. This architecture is important because it moves beyond counting how many people attended courses towards asking a much more meaningful question: how capable is the population at actually using AI?
Every country should begin asking the same question. AI literacy cannot become another divide between young and old, university graduates and non-graduates, executives and frontline workers, or large cities and smaller communities. A 19-year-old university student needs it, but so does a 55-year-old government employee, a nurse, teacher, farmer, lawyer, small-business owner and everyone enjoying their retirement. Not everyone needs to become an AI engineer, just as speaking English never required becoming an English professor, author or poet. What matters is having sufficient fluency to function confidently in the environment around you.
Governments once invested in roads, electricity, schools and broadband because these infrastructures enabled citizens to participate in the economy. AI capability is beginning to look like another form of national infrastructure, requiring universal foundational learning, affordable access to leading tools, workplace practice, sector-specific training and continuous assessment.
Twenty years from now, asking whether someone knows how to use AI may sound as strange as asking an office worker today whether they know how to use the internet. We have entered another country, its language is AI, and all of us need to learn to speak it.
