AI job risk checker: what 177 assessments reveal, and a new version launches today

The WINSS AI job risk checker collected 177 completed responses between March 24, 2025 and May 15, 2026, and the results place the typical respondent in the low-to-moderate exposure range: the median score was 71 out of a possible 160, and only 10 respondents (5.6%) landed in the highest risk band. On July 20, 2026, WINSS is launching an updated version of the tool that adds a live country comparison drawn from the OECD. This article reports what the first 177 assessments show, how the figures were produced, and why the new version changes what the tool can tell a user.

The checker is a 16-question self-assessment. Respondents rate aspects of their work on a 0-to-10 scale, where 0 marks work that is highly human, creative, or unpredictable and 10 marks work that is highly structured, repetitive, or automatable. The 16 answers are summed into a single score between 0 and 160, so a higher total signals higher self-reported exposure to automation. No registration is required and no personal data is collected.

How the AI job risk checker data were analysed

The dataset is a single export of 177 completed submissions from 169 unique respondents, collected through the Tally.so form that powers the tool. Eight respondents completed the assessment more than once; all 177 records are kept rather than reduced to one per respondent, because a repeat run may reflect a genuine re-assessment of a different role. Removing the eight repeat submissions does not change the reported mean or median.

Each record holds 16 item scores (0-10) and a computed total (0-160). The analysis reports the mean, median, standard deviation, and quartiles of the total; the mean of each of the 16 items; and the count of submissions per month. To describe the spread, responses are grouped into four bands at equal 40-point intervals: Very Low (0-40), Low-Moderate (41-80), Moderate-High (81-120), and High (121-160). These bands were defined for this analysis only; they are not part of the tool’s own scoring, and the tool applies no external weighting or occupational benchmark.

Source: WINSS AI job risk checker submission export, March 24, 2025 – May 15, 2026.

Score distribution: most results cluster in the low-to-moderate range

The distribution is roughly bell-shaped and slightly right-skewed. Most responses fall between 48 and 95, with a smaller group recording high scores. The Low-Moderate band (41-80) holds the largest share at 83 responses (46.9%), followed by Moderate-High (81-120) at 56 (31.6%). Twenty-eight respondents (15.8%) scored 40 or below, the profile of work with strong human or creative content, and 10 (5.6%) scored above 120.

The middle 50% of respondents scored between 51 and 88, an interquartile range of 37 points. The five-number summary below sets the quartiles and the observed range on the true 0-160 scale, with the mean of 72.2 marked just above the median of 71 – the small gap between them reflects the mild right skew.

Which work dimensions scored highest for AI exposure

Breaking the score into its 16 items shows where the exposure sits. The tool’s public description names 11 of the dimensions; the labels for five items (Q11 to Q15) are not individually published, so their means are reported as unlabelled. Because the total is a simple sum, the 16 item means add up to the mean total: the average item score is 4.51, just below the scale midpoint of 5.

Three dimensions pull scores toward the automatable end. Adaptability to technology records the highest mean at 5.86, followed by specialised knowledge at 5.55 and data-driven work at 5.40. At the other end, industry evolution has the lowest mean at 3.64, with creativity and innovation at 3.93 and human interaction at 4.03 close behind – the dimensions tied to work that is harder to automate.

The full item table, including the five unlabelled questions, is below.

What the pattern echoes in automation research

The low-scoring dimensions line up with the tasks that automation research has long identified as hard to computerise. In “The Future of Employment”, Carl Benedikt Frey and Michael Osborne (Oxford Martin School, 2013) estimated that 47% of United States employment sat in a high-risk category over the following one to two decades, and identified three engineering bottlenecks that hold automation back: perception and manipulation, creative intelligence, and social intelligence. The checker’s lowest means fall on the matching dimensions – physical tasks, creativity and innovation, and human interaction – so respondents are, in aggregate, locating their own defence against automation where that study locates it.

Occupation-level estimates that model tasks within jobs are more conservative. The OECD working paper “Automation, skills use and training” (Nedelkoska and Quintini, 2018) put about 14% of jobs across member countries at high risk of automation, with a further 32% likely to change substantially. That framing – more roles reshaped than removed – is a useful counterweight to any single self-assessment score. The WINSS checker is a self-report instrument and is not calibrated against these studies; the alignment is qualitative, not a validation.

Submission volume over 15 months

The tool launched on March 24, 2025 and drew 2 responses in its first week. Volume rose through the spring, peaked at 32 submissions in June 2025, and reached a second high of 28 in August 2025. From September 2025 the monthly count declined, settling at 1 to 4 per month through the first half of 2026 – the pattern of organic traffic to a self-contained tool page after its launch window.

What the numbers cannot tell you

The checker is a self-reported instrument. Scores reflect how respondents perceive their own work, not an independent classification of their occupation or employer-verified task data. Self-selection is built in: people who find and finish a tool on winssolutions.org are not a random sample of the workforce. The form is anonymous and collects no country, sector, age, or employment data, so the results cannot be broken down or generalised to any population.

Three design points also shape the numbers. The score is an unweighted sum, so every question counts equally toward the total; the four risk bands were set at equal 40-point intervals for this analysis and are not part of the tool’s own guidance; and the labels for five of the 16 questions are not published, so their contribution is visible in the total but not interpretable by dimension. With 177 responses, the figures describe this group of respondents and no one else. WINSS coverage such as “79 jobs that are AI-proof, for now” and “the 48 jobs AI will replace or impact between 2026 and 2030” offers occupation-level context that a personal score cannot.

An updated AI job risk checker launches today

On July 20, 2026, WINSS is launching an updated version of the checker, and the change answers the clearest limitation above: a score built only from self-perception. The new version keeps the self-assessment and adds an external, measured benchmark. A user selects their country and sees the share of businesses there that use AI, drawn live from the OECD “ICT Access and Usage by Businesses” database, next to the European Union and OECD averages.

That addition is why the new version is an improvement, on four concrete points. First, it places a subjective result beside an independent, source-backed statistic, so a personal score is read in context rather than in isolation. Second, it adds the country dimension the original never captured, since the anonymous form recorded no location. Third, the figures are retrieved at the moment of use through the OECD SDMX API and each is labelled with its data year, so the benchmark stays current as the OECD updates the dataset, and it breaks the number down by business size and by year rather than showing a single value. Fourth, the assessment is grouped into five dimensions drawn from named research – the Frey and Osborne bottlenecks, the OECD task-based approach, and the World Economic Forum “Future of Jobs Report 2025” – so the scoring rationale is explicit.

One point of precision matters. The OECD figure measures how widely businesses in a country have adopted AI; it is a context signal for how fast AI is entering workplaces, not a direct measure of any individual’s job loss. The WINSS AI job risk checker remains a guide for reflection, now with a verified national benchmark attached to the personal result.

About the WINSS AI job risk checker

The World Innovative Sustainable Solutions Association (WINSS), publisher of winssolutions.org, released the Free AI Career Risk Assessment Tool on March 24, 2025, within its Economy coverage of artificial intelligence and the labour market. The tool was created by Figen Sekin and built on the Tally.so form platform to be completed anonymously in a few minutes, scoring 16 dimensions of job exposure to AI on a 0-to-160 scale. It sits alongside WINSS reporting on AI and work, including the February 2026 article on how a “Belgian study found young workers among the first exposed to AI at work”. The 177 responses reported here cover the tool’s first 15 months, and the updated version launching on July 20, 2026 carries that work forward by pairing each self-assessment with live OECD adoption data.

Sources: WINSS Free AI Career Risk Assessment Tool; Oxford Martin School (Frey and Osborne); OECD; OECD Data Explorer

Featured image: photo by Felicity Tai on Pexels (free Pexels license).

Become a Sponsor

Our website is the heart of the mission of WINSS – it’s where we share updates, publish research, highlight community impact, and connect with supporters around the world. To keep this essential platform running, updated, and accessible, we rely on the generosity of you, who believe in our work.

We offer the option to sponsor monthly, or just once choosing the amount of your choice. If you run a company, please contact us via info@winssolutions.org.

• donate monthly
• donate once only

Select a Donation Option (USD)

$5.00 (p/m) $10.00 (p/m) $20.00 (p/m) Other (p/m)

Enter Donation Amount (USD)

Figen Sekin

I specialize in sustainability education, curriculum co-creation, and early-stage project strategy. At WINSS, I craft articles on sustainability, transformative AI, and related topics. When I’m not writing, you’ll find me chasing the perfect sushi roll, exploring cities around the globe, or unwinding with my dog Puffy — the world’s most loyal sidekick.