The jobs AI can do — and those it shouldn’t

Unlock the editor’s digest in free

FT editor Raula Khalaf selected his favorite stories in this weekly newsletter.

Generator AI is a converting technology that has the potential to define the nature of the work. Understanding its role in the workplace and it distinguishes from the past automation, which requires changes from AI Can Do it what do Should Tackle

The general analysis of the influence of the geni on the workers focus on whether the technology can do specific tasks. These national studies often break a task and evaluate the share of the works that can implement technology. For example, a call center includes conversation with customers in general jobs for customer service representatives, interaction recording and resolving anxiety or growing anxiety. Genny can handle these tasks, implies that it can displace these national staff.

But consider a profession that may initially appear equivalent: an emergency service phone operator. The two tasks share the same task. Should we expect that they will face the risk of equal levels of automation? The answer is more shorter than technical skills alone. Beyond moral considerations, these national roles automatically introduce the complex trade-offs associated with the economy, task design and operational inter-dependency.

We believe that four important questions of the organization should be considered when thinking about automation.

First, how complicated the work is? The complexity is both the main driver of human labor and AI expenditure. Emergency service sectors solve a variety of problems, associated with a complex that exceeds the repetitive interaction of the customer service representative. In general, the more complicated the work is, the less likely it is to be automatically, since people are for the present – better than the machines that manage the complications.

Second, how often is the work? The higher the frequency, the higher the chances of being automated. The machines have an obvious advantage to maintain speed during prolonged periods. Often repeated interactions with clients strengthen customer service representatives for AI replacement in the economic field.

Financial Times Global MBA Rankings

This article is from 2025 MBA Report And Ranking

Third, how much is the tasks connected? In the case of a service supply or product manufacture, many works are often involved in a chain of inter -associated work done by different workers and machines. What happens during the handoff in the work is often ignored. The handoff process shows the cost of dividing from inefficiency and defects.

The initial work for the Customer Service Representative involves a conversation with the customer, while the final work is solving their problem. When different workers or machines are involved, the handoff in this work can be expensive. If the final resolution worker does not interact with the customer initially, additional time will be required to review all the previously collected information.

Although high division expenditures are technically possible, companies should be discouraged from the work divided between people and generators AI. The call on the initial tridge on emergency services may seem expensive to automatically, but may lose significant information when converting a human transmitter from AI.

Fourth, what is the cost of failure when performing a task? Mistakes by emergency motivists create significant risks in particular life-or death circumstances. And Genny may be less precise than some of the past forms of automation.

These questions should be guided by companies considering automation and helping to explain why Zennie affects specific professions more than others. Consider the computer programmer for example. Extensive, well-enabled coding examples enable geny to provide effective solutions for complex tasks. Many coding work is a good fit with high frequency and repeated genius.

Prior to genius, programmers have divided large coding projects and reduced the cost of distributing innovations such as development platforms and modular designs. The safe test environment keeps the cost of failure low, as many errors can be detected cheaply in the geni-exposed code. In our structure, these features help to explain why programmers, automation’s traditions, are facing an increased disruption from the genius, beneficiaries.

The above four questions are highlighted that the generator makes AI as unique as automation technology. As it developed, Genny shows the ability to handle complex tasks at high speed, making it more versatile than the traditional thawing automation. By providing a pause non -interface and natural language processing capacity, genius progressively reduces the cost of refutation by comparing the traditional concessional automation. However, uncertainty surrounding the output of the genius potentially increases the risk of failure in any work.

Generator AI is a converter technology that is likely to re -shape labor markets. Its final effect and the possibility of acceptance are shaped by the work structure in a particular profession. The complexity of the work, their frequency, fragmented expenditure and cost of failure, taking together, spending costs and the balance between hidden expenditures affect the balance.