AI, Robotics and the Reality of Job Risk

There is no shortage of charts showing which jobs are “at risk” from AI. Most of them miss the point. They focus on what technology can do, not what organisations will actually implement.

The reality is more nuanced. They focus on what technology can do, not what will actually happen. In reality, job risk is shaped by four factors:

  • AI replaces thinking

  • Robotics replaces doing

  • Economics determines what is viable

  • Timeline determines when it happens

AI & robotics displacement risk

Adjusted risk score by role and industry — sorted high to low within each row

High risk Low risk
Bars = time to displacement (1=Long, 3=Short)
Metric:
Time horizon:

A Structured View of Job Risk

The model presented here moves beyond a single “risk score” and breaks job impact into the components that actually drive change. This matters because most roles are not simply “replaced” or “safe.” They sit somewhere in between, influenced by a combination of exposure, practicality, and timing. By separating these factors, the table is designed to reflect how roles are likely to evolve in practice, not just in theory.

How to Read This Model

Each role is assessed across two independent exposure dimensions.

  • AI Exposure - Measures how much of the role is driven by language, analysis, pattern recognition, or structured digital processes. These are areas where AI can already perform at scale, often with immediate cost advantages.

  • Robotics Exposure - Measures how much of the role involves physical, repeatable, and predictable tasks. These are areas where automation is possible, but often dependent on environment, standardisation, and capital investment.

These two dimensions are not interchangeable. A role can score highly in one and very low in the other.

From these, we derive:

  • Combined Risk - The higher of the two exposure scores. This reflects the maximum pressure on the role from either form of automation.

However, exposure alone does not determine outcomes.

From Capability to Reality

To move from theoretical capability to real-world impact, each role is assessed against Role Compression Constraint.

This captures the practical barriers to reducing or replacing the role, including:

  • cost of implementation

  • integration into existing systems and workflows

  • trust, accountability, and regulatory considerations

  • variability of the working environment

  • fragmentation of demand

In many cases, these constraints outweigh the benefits of automation, particularly where work is non-standardised or where errors carry significant consequences.

This allows us to calculate:

  • Adjusted Risk - A more realistic view of impact, reflecting what organisations are likely to implement rather than what is technically possible.

Introducing Time as a Core Dimension

Even where exposure is high and constraints are low, change does not happen immediately. Adoption takes time.

Organisations need to:

  • redesign workflows

  • integrate new systems

  • manage risk and accountability

  • align internally on how work is performed

As a result, disruption occurs in stages rather than in a single wave. To reflect this, each role is also assigned a Time Horizon:

  • Short term (0-3 years)
    Roles where the technology is already viable and the barriers to adoption are low. These tend to be digital, repeatable, and high-volume.

  • Medium term (3-7 years)
    Roles where capability exists, but adoption requires structural change, trust, or integration across systems.

  • Long term (7-15+ years)
    Roles with high practical constraints, often involving physical complexity or fragmented environments, where large-scale automation is slower to materialise.

Time is not simply a by-product of risk. It is a defining variable. Roles with similar exposure can evolve at very different speeds.

What This Reveals

When these elements are viewed together, several consistent patterns emerge.

First, exposure does not equal immediate disruption. Many roles that appear highly exposed remain stable in the near term due to cost, complexity, or organisational inertia.

Second, economic viability acts as the primary filter. Automation is adopted where there is a clear and immediate return, not simply where it is possible.

Third, change is uneven. Some roles compress quickly, particularly those that are digital and standardised. Others evolve gradually, even where exposure is high.

Finally, the impact is often internal to roles rather than across entire industries. Execution-heavy tasks are reduced, while oversight, judgment, and decision-making become more important.

Interpreting the Table

The table should not be read as a prediction of job loss. It is a structured view of pressure, feasibility, and timing.

Roles with:

  • high adjusted risk and short timelines are most likely to change quickly

  • high exposure but longer timelines are more likely to evolve gradually

  • low adjusted risk and long timelines are relatively more stable

The Key Takeaway

Understanding job risk requires more than a single score. It requires separating:

  • what can be automated

  • what is worth automating

  • and when that change is likely to occur

Most roles are affected in some way. Very few are removed outright. The real shift is not whether work disappears, but how it is redistributed.