Bill Gates helped build the personal-computer revolution. Now he is warning that artificial intelligence could produce a far greater upheaval—and governments may be dangerously unprepared for what follows.
Bill Gates has spent much of his life on the winning side of technological change.
As Microsoft’s co-founder, he helped take the personal computer from a specialised machine used by a relatively small number of people and turn it into an indispensable part of modern life. The revolution transformed offices, businesses and entire industries, created millions of new jobs and made Gates one of the world’s wealthiest men.
Now he is warning that the next technological revolution could be considerably more disruptive.
In a recent essay, Gates argued that artificial intelligence is likely to transform the labour market on a scale that will be difficult for governments and workers to absorb. Unlike earlier technological changes, which unfolded over decades and gave societies considerable time to adjust, he expects AI to move rapidly into professions ranging from law and medicine to customer service, software development and manufacturing.
The warning deserves attention partly because it comes from someone who knows what technological disruption looks like from the inside.
But Gates is not predicting the end of work, nor is he arguing that artificial intelligence should be stopped. He continues to see enormous potential for AI in medicine, education, scientific research, agriculture and public services.
His concern is what happens to people during the transition.
That question is already becoming harder to avoid.
When the machine becomes cheaper than the worker
The first industrial revolutions replaced forms of physical labour. The computer revolution automated calculations, record-keeping and other routine tasks.
AI is moving further into work that was once considered distinctly human.
A junior lawyer researching cases, an accountant preparing routine reports, a programmer writing standard code, a customer-service employee answering predictable questions and a medical professional processing information may increasingly find that some of their traditional tasks can be performed by software.
For companies, the attraction is obvious. If a machine can perform a task more quickly, continuously and at a lower cost, the commercial incentive to use it will be difficult to resist.
That creates a problem that goes beyond the immediate loss of individual jobs.
Many professions depend on junior employees doing routine work before acquiring the experience needed to take on more complicated responsibilities. The young lawyer learns by researching cases. The trainee accountant learns by preparing accounts. The junior programmer learns by writing and debugging code.
If AI takes over much of that work, where does the next generation acquire its professional experience?
It is a question that has received considerably less attention than the headline prediction that AI will create new jobs.
The answer cannot simply be that displaced workers should “learn AI”. Someone still has to determine which skills will remain valuable, how people will acquire them and who will pay for the transition.
The people entering the workforce may feel it first
That problem could be especially serious for countries with large and youthful populations.
Nigeria is one of them.
Every year, hundreds of thousands of young Nigerians enter a labour market already struggling to absorb graduates. Many find work in sectors where routine administrative, analytical and customer-service functions are precisely the areas in which AI is advancing fastest.
The technology also offers Nigeria enormous opportunities. It could make businesses more productive, extend access to medical expertise, improve agricultural decision-making, assist teachers and reduce the cost of delivering public services.
But productivity and employment are not the same thing.
A company that previously required twenty employees to perform a particular category of work may eventually need five people equipped with powerful AI systems.
For a mature economy with strong social protection and extensive retraining programmes, such a transition would still be difficult.
For a country where many workers have little financial cushion and where public systems for retraining and unemployment support remain limited, the consequences could be considerably harsher.
Gates is asking an uncomfortable question about robots
The debate has so far focused heavily on white-collar employment. Gates believes that will not remain the whole story.
As AI becomes more capable of controlling machines and robots, occupations involving physical work could also face increasing pressure. Construction and hospitality are among the areas he has identified as potentially vulnerable to advances in robotics.
That possibility changes the scale of the problem.
If technology eliminates some routine office jobs while creating new technical occupations, workers may be able to move from one category to another.
But if increasingly capable machines begin performing both cognitive and physical tasks, the assumption that technological progress will automatically generate enough replacement employment becomes much harder to defend.
History offers reasons for optimism, but it does not guarantee that every worker displaced by one technological revolution will find a suitable place in the next one.
Should some work remain human?
One of Gates’s more provocative suggestions is that certain occupations or functions should remain reserved for people even where machines become technically capable of performing them.
Healthcare provides an obvious example.
An AI system may eventually analyse medical information more rapidly than a doctor. That does not necessarily mean patients will want machines to perform every human function associated with medicine.
There is a difference between diagnosing a disease and sitting with a frightened patient.
There is a difference between processing information and teaching a child.
There is a difference between producing a legal document and exercising professional judgment on behalf of someone whose liberty, property or livelihood is at stake.
The debate therefore cannot be reduced to what machines can do.
It must also ask what society believes machines should do.
And who pays for the transition?
Gates has also revived an idea that is certain to provoke fierce debate: taxing AI-driven automation, including robots, to compensate for some of the public revenue that may disappear when human employment declines.
The proceeds, he argues, could help finance retraining and other social programmes for workers affected by automation.
Technology companies are likely to object that such measures could discourage innovation and investment.
There is nevertheless a serious fiscal question beneath the argument.
Governments depend heavily on economic activity generated by human labour. Workers pay income taxes and social contributions; businesses pay taxes partly from the economic activity created by their employees.
What happens to that system when a substantial portion of the work is performed by machines?
The answer cannot be left entirely to technology companies to determine.
The legal profession has its own problem
Gates’s inclusion of law among the professions likely to be transformed by AI should command particular attention from the Nigerian Bar.
The question is no longer whether artificial intelligence will be used in legal practice. It already is.
Legal research, document review, transcription, drafting, case management and other routine functions are increasingly being assisted by AI tools.
The difficult questions concern what follows.
If a machine produces a defective legal argument, who bears responsibility?
If AI allows one lawyer to perform work that once required several junior lawyers, how will young practitioners acquire experience?
Will clients expect lower fees because technology reduces the amount of time required for certain tasks?
And if junior lawyers no longer spend long hours doing routine research and drafting, what replaces that apprenticeship?
These are not distant questions. They belong in the conversation about the future of the profession now.
Nigeria cannot afford to treat this as somebody else’s problem
The temptation for developing countries is to regard AI primarily as something produced elsewhere and consumed here. That would be a mistake.
Nigeria has much to gain from the technology, but it also has a great deal to lose if the transition is poorly managed.
Government policy will have to address education, taxation, employment, professional regulation, data protection and social protection at the same time.
Schools will have to teach young people skills that machines cannot easily replicate, while also ensuring that students understand how to work effectively with AI.
Universities will have to reconsider curricula that prepare graduates for occupations in which some routine functions may soon be automated.
Professional bodies will have to rethink training and entry-level practice.
Employers will have to decide whether AI is being deployed simply to reduce headcount or to make workers more productive.
And government will have to confront the possibility that the traditional relationship between economic growth and job creation may become less predictable.
The warning from the man who helped start the revolution
There is an irony in Gates’s position that should not be missed.
The personal computer did not destroy work. It changed the nature of work and created an enormous new economy around it.
AI could do something similar, but on a much larger scale and at a much faster pace.
A worker cannot retrain overnight. A university cannot redesign its curriculum every six months. A government cannot build a social-protection system at the speed at which a software model improves.
Technology may move in months.
Institutions move in years.
That is the gap governments now have to close.
The sensible response is neither panic nor blind enthusiasm. Artificial intelligence could produce extraordinary gains in medicine, education, science and productivity. It could also widen inequality and leave millions struggling to find a place in an economy that suddenly requires fewer people to perform familiar jobs.
The choices made now will determine which of those futures becomes more likely.
Bill Gates knows something about technological revolutions. He helped build one.
His warning is therefore worth considering for what it is—not a prediction that human beings will become obsolete, but an argument that the institutions responsible for preparing human beings for the next economy cannot afford to wait until the machines have already arrived.
