Automation risk scores are not all created equal. The same occupation can receive a 35% risk score from one research group and a 78% score from another, and both can be technically defensible. The difference comes down to methodology: whether the research measures whole jobs or individual tasks, which AI capabilities it assumes, and how far into the future it projects. Understanding what these scores actually measure is essential before using them to make decisions about your career. Task-level analysis consistently produces more useful results than occupation-level analysis, because two people with the same job title can face dramatically different exposure depending on how their actual work is structured.
What the research says
92M
jobs displaced by 2030
WEF Future of Jobs 2025
170M
new roles created by 2030
WEF Future of Jobs 2025
41%
of employers plan AI-driven headcount reductions
WEF 2025
55K
job cuts explicitly attributed to AI in 2024
Challenger, Gray and Christmas
Automation risk scores measure the share of tasks within an occupation that current or near-future AI systems can perform at human level or above. The original Oxford Martin School methodology looked at 702 occupations and rated them on nine technical capabilities: fine motor skills, creative intelligence, social intelligence, and six more. More recent O*NET-based scores use task-level data from the US Department of Labor and weight each task by how much time workers spend on it.
Because the same job title often covers very different actual tasks in different organisations. A marketing manager at a small firm may write all their own copy, manage ad campaigns, and produce reports. A marketing manager at a large company may do none of those tasks and instead manage a team, set strategy, and present to executives. The first version has a much higher automation exposure than the second, even though both share a job title.
The most rigorous publicly available sources are: O*NET from the US Department of Labor (granular task-level data, updated regularly), the World Economic Forum Future of Jobs 2025 report (employer survey data from 1,000+ companies), and McKinsey Global Institute research on task automation across 60 economies. The original Oxford Martin School 47% figure is now considered outdated because it pre-dates the large language model era and did not account for task-level nuance.
You can look up your occupation in O*NET Online at onetonline.org, which provides detailed task breakdowns and technology-related data. The limitation of lookup tables is they give you the average for your occupation category, not your specific role configuration. A personalised assessment that accounts for your actual daily tasks gives a more accurate result than any lookup.
Not directly. A high automation risk score means a high proportion of your tasks are technically automatable. Whether employers act on that depends on AI adoption velocity in your industry, the cost of implementation, regulatory constraints, and labour market conditions. The OECD estimates 14% of jobs are highly automatable right now, but actual displacement is running well below that because adoption takes time. The risk score is a leading indicator, not an immediate verdict.
Free Assessment
Free Fossil Score assessment. No account required.
Get My Fossil Score