By Stephen Milner · UtilityForge · Last reviewed: May 2026
This tool draws on the most widely cited academic research into automation risk to give you a personalised, occupation-level view of how AI is likely to affect your working life. The goal is not to alarm you but to give you enough information to make good decisions about skills, training, and career direction.
The automation probability scores come primarily from the Oxford Martin School's landmark 2013 paper "The Future of Employment" by Carl Benedikt Frey and Michael Osborne, which examined 702 occupations and estimated the share of their tasks that could be performed by a machine given advances in AI and robotics. That work has since been updated and refined by the World Economic Forum's Future of Jobs reports and McKinsey Global Institute research on task-level automation potential.
For occupations in technology and creative fields, the probabilities have been adjusted to reflect the material impact of generative AI tools that arrived after the original study was published. Content writing, graphic design, and coding assistance are meaningfully higher risk today than the 2013 figures suggested.
A score of 0.94, for example, does not mean 94% of people in that role will lose their jobs within a specific year. It means that 94% of the tasks typically performed in that role are, in principle, automatable using current or near-term AI. The actual pace of displacement depends on the cost of automation relative to labour, regulation, public acceptance, and the degree to which new tasks emerge alongside automation.
Roles with low scores are not immune from change. Even a surgeon (0.4%) still benefits from AI-assisted imaging and robotic surgery tools. The difference is that for low-risk roles, AI acts as an amplifier rather than a replacement.
The pattern is consistent across studies. Roles built on a narrow set of well-defined, repeatable tasks are most vulnerable. Data entry, invoice processing, and standard call handling are examples of work that follows clear rules and can be described precisely enough for a machine to perform.
Roles requiring unpredictable physical environments, nuanced human relationships, novel problem-solving, or complex ethical judgement score much lower. A plumber navigating an unfamiliar building layout, a therapist reading emotional cues in a session, or a lawyer constructing a novel legal argument all involve dimensions that current AI systems cannot reliably replicate.
If your occupation shows a high score, that does not mean your skills are worthless. It means the tasks are changing, and the most valuable version of your role is likely to shift toward the work that automation cannot do: advising, supervising automated systems, managing exceptions, building client relationships, and exercising judgement in complex situations.
The career pivot suggestions in this tool are chosen for skills transferability, not just lower risk. Every suggested role shares meaningful overlap with what you already know how to do, which shortens the distance between where you are now and where the demand is heading.
Does a high score mean I will lose my job?
Not automatically. The score reflects how many of your role's tasks could be performed by a machine, not a prediction of when or whether your specific employer will automate them. Cost, regulation, and social factors all shape how quickly automation actually reaches a given workplace. The score is best read as a signal for where to focus skill development rather than a countdown.
My job title is not in the database. What should I try?
Search for the closest match by function. If you are a "Senior Credit Analyst", try "Financial Analyst". If you are a "Recruitment Manager", try "HR Manager". Most roles sit within a broader occupational category. The database covers 65 occupations across 14 sectors and will grow over time.
Why is the score for my job different from what I have seen elsewhere?
Different studies use different methodologies. Some assess whole occupations, others break roles into individual tasks. This tool uses the Oxford Martin School task-level approach, adjusted for post-2022 generative AI developments. Graphic designers and content writers, for example, score higher here than in the original 2013 study because large language models and image generation tools have materially changed the task landscape since then.
Does this tool know what AI can actually do today?
The base data comes from academic research, supplemented with judgements from WEF and McKinsey analyses published through 2024. The tool does not query live AI systems or reflect the capabilities of any specific product. For rapidly moving fields like software engineering, legal research, and medical imaging, treat the scores as indicative rather than precise.
Are some industries being automated faster than others?
Yes. Finance, administration, and logistics are seeing the most active deployment of automation tools right now because the tasks are well-defined and the volumes justify investment. Healthcare and education are moving more slowly, partly due to regulation and partly because the human relationship is central to the service. Skilled trades are constrained by the cost of physical robotics, which remains high relative to human labour in most markets.
My role showed as low risk. Does that mean I do not need to worry?
Low risk does not mean no change. Even in roles rated below 10%, AI tools will increasingly handle routine sub-tasks like documentation, scheduling, and first-draft outputs. The opportunity in low-risk roles is to use those tools to do more work at higher quality, rather than to ignore them. The risk of not engaging with AI tools is becoming its own career risk regardless of automation probability.
What should I do if I want to move into a lower-risk career?
Start with the pivot suggestions on your result. They are chosen because the skill overlap is genuine, not just because the risk score is lower. Identify the gap, find the shortest training path to fill it, and look for lateral moves within your current employer before making a full switch. Most career transitions in this context are steps rather than leaps.
How often is the data updated?
The underlying occupation probability scores are updated when major new research is published, typically annually. The tool reflects the state of knowledge as of early 2026.