f you are in the top one percent in the world in your skillset or area of expertise, you are probably not at the risk of being replaced by artificial intelligence. That is probably the only thing one can say with a degree of confidence about the future of work (and skills) in the Age of AI.

OpenAI, the non-profit that ushered in the Generative AI “revolution” when it launched a chatbot based on its large language model technology, estimated whether (and to what extent) AI would replace human (skills). It concluded that AI would replace about 10 percent of the tasks of 80 percent of the people and 50 percent of the tasks of 19 percent of the people.

This means, according to OpenAI, that artificial intelligence could make 19 percent of the workers ‘half’ unemployed and substantially automate some work of the vast majority. I believe that these are conservative estimates.

In more recent months, many technology pundits—such as Bill Gates—have argued that the AI will bring about a massive disruption to the world of work, creating unemployment amongst a vast number of people, making the need for retraining tens of millions of people. Gates goes as far as to say that the AI would spare only 3 to 5 professions and make the rest redundant.

For a country like ours, OpenAI’s estimates could easily reverse. 50 percent of the tasks of 80 percent of the people could be automated and 10 percent of the next 19 percent. This means that for certain types of professions—mostly knowledge workers—80 percent of the people could become ‘half’ unemployed. The reason for this is that the vast majority of Pakistanis—as many as, perhaps, 90 percent—usually perform mid-to-low level knowledge work. There are content writers, designers, coders-developers, accountants, lawyers and teachers etc, who are not performing at the cutting-edge but rather doing things that machines can easily do. They are replaceable. Many of these people are at the risk of being replaced by machines—or more plausibly other people working with machines.

Take content writing, graphic designing or social media marketing, for instance. Today, with one prompt, or a master design, AI can easily create tens, hundreds of copies of varying text for a blog, website or social media. Earlier there used to be teams of people trying to create these variations or different adaptations. Today it can easily be done, even arguably better, by someone using the AI.

This has major implications for the future of work and skills, especially for the young.

It is undeniable that the AI is here to stay and whether or not we are able to progress to the next frontier of Artificial General Intelligence (AGI)—meaning we will be able to create machines that totally replace humans and could even threaten the existence of our species—the AI that we have today is better than humans in performing many mundane and repetitive tasks.

Second, the only way to counter the resulting disruption to skills and professions is to learn how to use AI to supplement what you already do and try to use it to do it better. In the near future, we will have one person using AI to do the job of three to four (or more) people. We are likely to see improvements in productivity for those who embrace AI and learn to use it effectively and de-employment or mass disruption amongst those left behind.

Third, the AI will also, if properly used, create opportunities for us to enhance the quality and intellectual value of our work. While initial studies in the uses of AI—particularly by teachers and students—suggest that AI leads to learning losses and cognitive short-circuiting as users tend to begin to depend too much on AI outputs, its effective use requires people to learn how to use it to supplement their own knowledge (and work) rather than mindlessly replacing it.

Using AI effectively brings us back to the core idea of critical thinking which was and has always been the core purpose of education, but something that our education systems have struggled with.

Take computer programming. There is a lot of discussion today on whether AI can programme as well as humans. There is a near consensus that while AI cannot programme as well as the best programmers but for ordinary coding tasks it can, at least as well as our average coder. Taking cue from this, big tech companies such as Microsoft and Amazon, are experimenting using AI (humans + AI combinations) to replace programmers. With hundreds of thousands of programmes on their payroll, even a 15 to 20 percent reduction could means tens of thousands of job losses. Those who will survive will focus on higher level work such as architecture, design, algorithms, checking-testing, etc leaving the everyday mundane coding to machines.

Even in other areas and professions, those who can learn to use AI properly will be able to focus on higher order thinking and work rather than mundane and repetitive tasks. Teachers can now focus on critical thinking rather than checking the language of their learners.

In fact, using AI effectively brings us back to the core idea of critical thinking which was and has always been the core purpose of education, but something that our education systems have struggled with.

In the early days of Generative AI, there was a sudden craze to learn “prompt” engineering which was a fancy name given to how one asks questions so that one could optimise the way the AI answers. Part of the prompt engineering has to do with understanding how the AI and LLMS work and based on that, learning a few tricks to trigger certain elements in its response. But a large part of prompt engineering is learning to ask the right questions.

What prompt engineering teaches us is to take a response by the AI, absorb it, and then come up with a more intelligent question. This is another name for critical thinking, really. What else is critical thinking but to ask questions and question answers.

So, the AI has brought us back to the basics. We need to learn critical thinking to be able to do well and prosper in the age of AI. The good thing is that now we have the necessary tools to do this. For several decades, we have been talking about personalised learning but we did not have the tools for it. With Generative AI and LLMs we now have the tools necessary to serve everyone, helping us learn what we want to and at the pace we want.

Personalised learning might ultimately be the killer app for education, provided we can find the solution to the behavioural challenge of motivating people to learn themselves.

The AI, even in its currently imperfect form will make a significant dent to the world of work and skills. If used effectively, it could liberate most of us from many tasks that we find mundane and repetitive and free our minds and hours to do higher order things.

Another emerging area is that of AI agents and Agentic AI where one can use the AI to do a number of tasks and make decisions on the way, almost acting like a human. The AI agents will be able very quickly to run entire social media campaigns, make itineraries and bookings as travel agents, act as salespersons or purchasers.

There are already one-person companies with a founder working with 5 to 50 AI agents as employees building products, pushing and selling those to the market.

This creates a world of possibilities for people who want to build businesses at low cost, perhaps even improving their chances of success.

There is now a clear imperative to teach people—all people, not just the young, irrespective of their professions — AI skills that will help them use it more effectively to enhance productivity and quality of their work.

We must act on our own to master AI skills to become better as professionals and be amongst those (human + AI) who would be doing the work of three rather than those whose work is made redundant.

The time to act is now.

The writer is a former member of the Planning Commission. He is currently the managing partner of INNOVentures Global (Pvt) Ltd and co-founder of PakGPT, a platform to promote AI skills and education. He is accessible at athar.osama@gmail.com.