AI job risk is skill gap not replacement
Most discussions about AI and jobs start with the same question: Will it replace us? That question is comfortable because it is binary. It offers a yes or no, a clear winner or loser. But it is the wrong question. It assumes the economy is a zero-sum game where tasks are simply reassigned from humans to machines. Cisco President Jeetu Patel has argued that this framing misses the point. He claims early data suggests AI will create more jobs than it eliminates over the next five years The Silicon Review. Whether he is right or wrong, the more interesting question is not about replacement. It is about what happens to the people who cannot keep up.
The Feedback Loop Between Speed and Skill
Patel argues that when AI makes one process faster, businesses do not stop. They discover new challenges, create new products, and explore new markets The Silicon Review. This sounds optimistic, but it hides a structural problem. The feedback loop between system deployment and system speed is not neutral Faster processes do not automatically create better jobs. They create different jobs. And those different jobs require different skills. The person who used to sort data by hand cannot simply step into a role that now requires designing the prompts that sort the data. The gap between the old job and the new one is not a step. It is a chasm. Patel himself acknowledges this when he says the bigger risk is not AI replacing people, but people failing to develop the skills needed to work with AI The Silicon Review. The feedback loop rewards those who can adapt fastest. For everyone else, it becomes a trap.
Patel paints a picture where AI-fluent workers become dramatically more productive. He claims they could be 50 times or even 100 times more effective in some forms of work The Silicon Review. That number is staggering. But consider what it means for the person who is not AI-fluent. If a handful of workers can do what a hundred used to do, the other ninety-five do not simply vanish. They remain in the economy, but their labor is now worth less. The feedback loop amplifies inequality. The system gets faster, the skilled get richer, and the unskilled get left behind. This is not a prediction of doom. It is a description of how every previous automation wave has worked. The difference this time is the speed. Previous waves took decades. This one is happening in years.
The Historical Pattern of Displacement
History offers a clear pattern. Every major technological shift has eliminated some jobs while creating others. The industrial revolution destroyed the livelihoods of skilled artisans who spent years learning a craft. It created factory jobs that required little training but offered steady wages. The information revolution eliminated typing pools and switchboard operators. It created roles for software engineers and database administrators. In both cases, the transition was brutal for the people caught in the middle. The new jobs did not appear overnight. They took a generation to emerge. And they often required skills that the displaced workers did not have.
Patel argues that protecting jobs by slowing AI progress is not the answer The Silicon Review. He is right about that. Slowing progress does not preserve jobs. It only delays the inevitable. But his alternative — focusing on AI education and reskilling — assumes that training programs can keep pace with the technology. That assumption is fragile. The half-life of a technical skill is shrinking. A course in prompt engineering that is cutting-edge today might be obsolete in two years. The feedback loop between system deployment and system speed does not just reward the skilled It rewards those who can learn continuously. And continuous learning is not something that a single training program can provide.
Patel highlights AI fluency as a defining advantage of the future workforce The Silicon Review. He is right. But fluency is not a binary state. It is a spectrum. On one end, you have the person who can use a chatbot to draft an email. On the other, you have the person who can build and fine-tune the model. The distance between those two points is enormous. And the jobs that get created will cluster toward the high end of the spectrum. The low-end fluency tasks will be automated themselves. This is the trap that Patel’s vision does not fully address. The feedback loop does not just create jobs. It creates a hierarchy of jobs. And the people at the bottom of that hierarchy are the most vulnerable.
The Question of the Goal We Pursue
Patel believes AI automation will create more jobs than it eliminates over the next five years The Silicon Review. Maybe he is right. Maybe the early data he mentions will hold up. But even if he is right, the number of jobs is not the only metric that matters. The quality of those jobs matters. The distribution of those jobs matters. The speed at which they appear matters. If AI creates a million high-skill jobs but eliminates two million low-skill jobs, the net number is positive, but the human cost is devastating. The feedback loop between system deployment and system speed does not care about human cost It only cares about efficiency.
Patel says that when AI makes one process faster, businesses discover new challenges and explore new markets The Silicon Review. That is true. But it assumes that the new challenges and new markets will generate jobs that are accessible to the people who lost their old ones. That assumption is generous. The new challenges that AI reveals are often technical. They require understanding of the system itself. The new markets that emerge are often digital. They require skills that are not evenly distributed across the population. The feedback loop does not automatically create opportunities for everyone. It creates opportunities for those who are already positioned to take advantage of them.
The question we should be asking is not whether AI will create more jobs than it eliminates. The question is whether the goal we pursue is the right one. If the goal is simply to maximize efficiency and productivity, then the feedback loop will run its course. Some people will thrive. Many will struggle. The system will get faster. The inequality will grow. But if the goal is to build a society where the benefits of AI are broadly shared, then we need a different approach. We need to think about the feedback loop not as a force of nature, but as a system we can design. We can choose to slow down the deployment in certain areas. We can choose to invest in social safety nets. We can choose to create jobs that are not purely about efficiency.
Patel argues that companies, governments, and workers must focus on AI education and reskilling The Silicon Review. That is a good start. But it is not enough. Education and reskilling are supply-side solutions. They try to make workers fit the system. They do not try to make the system fit the workers. The feedback loop between system deployment and system speed is powerful It will continue to accelerate. The only way to ensure that it does not leave millions behind is to intervene on the demand side. We need to create demand for the kinds of work that humans do well — the work that requires empathy, creativity, and judgment. AI can augment that work. It cannot replace it.
Patel is right that AI will not simply destroy jobs. It will transform them. But transformation is not the same as salvation. The feedback loop between system deployment and system speed is already in motion. The question is not whether we can stop it. We cannot. The question is whether we can steer it. The answer to that question will determine whether AI becomes a tool that lifts everyone or a force that divides us further. The wrong question is about replacement. The right question is about distribution. And that question has no easy answer.