These tasks include recognizing images, understanding language, making decisions, learning from data, and sometimes acting in complex environments. “Automation” more generally refers to machines or software systems taking over tasks that humans do—whether repetitive physical tasks (such as in factories), or cognitive or decision-making tasks (such as bookkeeping, data entry, or even composing reports). AI is a more advanced form of automation, because it can adapt, learn, generalize, and sometimes perform in uncertain or changing environments.

Concern about machines replacing human labor is not new. Each major technological revolution—mechanization during the Industrial Revolution, electrification, the rise of computers—has threatened some kinds of jobs even as it creates others. For example, when tractors replaced manual plows, many farmhands lost jobs, but demand for factory workers, toolmakers, transportation workers, and other roles increased. When computers came in, many clerical, bookkeeping, and data entry jobs changed or vanished, but new jobs in software, support, operations and information management proliferated.

So history shows two important lessons: that technology tends to change the kinds of jobs available rather than eliminate work altogether; and that transitions can be painful and uneven. Some people or communities suffer job losses or economic disruption while others benefit.

AI and modern automation are different in kind from past machines. They threaten both manual, repetitive tasks as well as routine cognitive tasks. Here are five areas that are most vulnerable:

Repetitive manual tasks: assembly line work, basic manufacturing, packing, sorting, etc. Robots and automated systems are already replacing or augmenting human work in many factories.

Routine data tasks and clerical jobs: bookkeeping, basic accounting, data entry, administrative tasks, routine customer service, etc. For example, chat-bots and automated call systems can handle many standard customer queries; software can automatically process invoices, receipts, and other financial documents.

Some middle-skill tasks: tasks that follow predictable rules but require judgment, such as basic legal review, proofreading, simple medical diagnosis, and even journalism (e.g., news summarization, sports reporting). AI tools are increasingly good at pattern recognition, data summarization, and even generating text.

Jobs with high redundancy or standardization: transport (e.g. truck-driving, taxi-driving with autonomous vehicles), retail checkout (self-checkouts) and warehousing (robotic pickers). However, jobs that are less predictable and require creativity, empathy, complex decision-making, physical adaptability, or intricate social interaction are likely to be more protected—at least for the time being. These include creative professions (artists, writers, designers), high-level management, caregiving, healthcare providers in many roles, roles involving negotiation or persuasion, and professions that require deep domain expertise.

While AI does threaten certain jobs, it also offers significant opportunities. Some workers might lose their traditional tasks, but gain new or transformed roles. Several trends are worth noting: Augmentation rather than replacement: In many sectors, AI is being used to assist human workers, making them more efficient rather than replacing them entirely. These include:

• Doctors using AI diagnostics tools to help them detect disease earlier;
• Journalists using AI to sift through data or produce first drafts;
• Lawyers using AI to search precedents or do contract review;
• New job creation: AI creates demand for workers who build, maintain, train, supervise, or audit AI systems. There are roles in data science, AI ethics, algorithmic fairness, cyber-security and others.

 Opportunities for efficiency and productivity: AI can reduce costs, enhance output, open new markets, and enable innovation. That can lead to economic growth, which in turn can produce more jobs

Whether AI is coming for your job depends on a variety of factors such as your field, your skills, your ability to adapt and external circumstances, such as policy and economic conditions. Here are some determinants: 

Task structure—How repetitive, routine, or well defined are the tasks? If much of your job involves predictable, repetitive work that can be codified, it’s more likely to be automated. However, if your job requires ambiguity, creativity, interpersonal connections, unpredictable physical environments, or ethical judgment, it’s much safer.

But whether we like it or not, AI isn’t coming, its here.

About the author: Michael J. Paton is a portfolio manager at Tocqueville Asset Management L.P. He can be reached at 212-698-0800 or by email at MPaton@tocqueville.com.

Published: October 16, 2025.