Before deployment · Exploratory short form

Assess an AI project before deployment.

A guided 18-item pathway for examining the project through the six factors in the Gollac report and three AI-related dimensions.

Start the assessment ↓Understand the method

18QuestionsTwo items for each of the nine dimensions.

≈ 10 minIndicative durationAnswer based on how the system is actually expected to operate.

LocalNo data sentAnswers do not leave the browser.

ExploratoryNot psychometricNo diagnosis, certification or scientific threshold.

01 · PrepareEssential guidance

02 · DescribeContext and use

03 · AssessOne question at a time

04 · InterpretProfile and discussion points

Initial assessmentNo identifying data

Answer for the use that is actually planned.

Use the most exposed population and the most demanding reasonably foreseeable situation. Do not rely only on the provider’s demonstration.

01A specific use

Choose a clearly defined task, population and organisation.

02Missing information remains visible

Choose “NI” rather than a middle response when the evidence is insufficient.

03A collective discussion

Then compare the result with the views of the people concerned.

Tool status

This short form is empirical and has not been scientifically validated. It is not a diagnosis, proof of compliance or authorisation to deploy.

Step 02Non-identifying context

Describe the project without naming it.

This information provides context for the profile. Do not enter any company, person, software or provider name.

Main sector

Approximate organisation size

Project stage

Population directly concerned

Assessment dateThis date will appear in the report.

Planned nature of use

Task or process concernedDescribe the function, not the product name.

What will the AI be used for? Select all that apply.

Generative assistant

Work organisation

HR decisions

Monitoring or control

Health or safety

Physical system

Step 03Answers stored locally

Question 1 of 18

Answer with the most demanding reasonably foreseeable situation for the population concerned in mind.

0Not at all true

1Slightly true

2Moderately true

3Largely true

4Completely true

NIInsufficient information

N/ANot applicable

A higher number describes a less favourable situation. NI and N/A are not converted to numeric values.

Exploratory resultReport generated locally

Descriptive project profile.

The results organise the information entered. They prove neither the presence nor the absence of risk, and cannot be used to compare people or organisations.

Why no overall score?

This short form has not been psychometrically validated and no scientific threshold has been established. A single average would hide specific mechanisms and missing information.

Nine separate dimensions

Descriptive mean · 0 to 4

Dimensions to discuss first

Relative ranking · not a threshold

Items scored 4

Examine independently from the means

Method & limits

A discussion aid, not a scientific measure.

The short form retains two items per dimension from the development long form. This reduction supports initial screening but also reduces coverage of each dimension.

01Nine dimensions shown separately

Six Gollac factors and three exploratory AI-specific mechanisms.

02No diagnosis

The result replaces neither work analysis nor clinical or legal expertise.

03No individual ranking

The questionnaire concerns a project and work situation, never an individual worker.

04No decision threshold

Continuing or stopping a project requires documented discussion, not an automated score.

Read the methodological overview →Is the system already deployed?

Supported use

Would you like to use the tool in an organisation?

Employers, HR teams, employee representatives, OHS services and research teams can get in touch to discuss scope, participants and conditions for using the short or long form.

Contact Dr Charles Broutin ↗