AI transparency
We use AI to help with tutoring, reading CVs, marking one written answer in the skill check, summarising evidence for staff and drafting job adverts. AI never ranks candidates for employers, never rejects an application and never awards the Vetted badge. A person reviews any result on request.
Last updated 28 October 2026
In short
- The table below lists every AI feature, what it sees and who provides it.
- No AI ranking, no AI rejection, no inferring protected characteristics.
- The skill-check scenario is the one AI mark that can matter, so a person reviews it on request.
- Your data is not used to train AI models.
1. Where we use AI
This is every place on the site where an AI model reads something you or an employer wrote. Where a provider is named, it is a processor working for us under contract. See the privacy notice for lawful bases, transfers and retention.
| Feature | What it does | Data it sees | Provider and place | Human oversight |
|---|---|---|---|---|
| AI tutor in lessons | Answers questions about the lesson you are on. | What you type, and the lesson context. | Anthropic (Claude), USA. If it is unavailable, other providers named in the privacy notice. | It makes no decision about you. We do not keep tutor conversations on our servers. The last few messages stay in your own browser until you start a new topic. |
| Course advisor on course pages | Answers questions about our courses. | What you type, with email addresses and phone numbers removed first. | Anthropic (Claude), USA, or the fallback providers in the privacy notice. | It makes no decision about you. Do not put personal details in the chat. |
| EDWartens AI in the student portal, the employer portal and the admin console | Answers questions about the page you have open and about EDWartens: what a page does, what to do next, how vetting, skill checks, verification and posting a job work. It points you to pages and never takes an action for you. | What you type, with email addresses and phone numbers removed first, the page you are on, and a few counts from your own account: a learner's stage and number of courses and applications, an employer's live jobs and new applicants, the number of items waiting for the owner and admins. Not your CV, your answers or your messages. | Anthropic (Claude), USA. If it is unavailable, other providers named in the privacy notice that do not train on prompts. | It makes no decision about you, never ranks or scores anyone, and cannot change anything. It is switched off during skill checks, assessments and job simulations. We do not keep the conversation on our servers: it stays in your browser tab and is gone when you reload or start a new chat. We log that a question was asked, with no text. |
| AI colleagues and marking in job simulations | Chat as fictional colleagues, and give a first-pass mark on written answers against a marking guide. | Your chat messages and written answers in the simulation. | Anthropic (Claude), and for chat Groq, Cerebras and NVIDIA, USA. | Marks are practice feedback with no legal or similarly significant effect. You can ask for any mark to be reviewed by a person. |
| First-pass marking of module projects | Gives a first mark on a graded project in the Siemens and Allen-Bradley Modules, against the published rubric. | The files you submit for that project. | Anthropic (Claude), USA. | Staff see every AI mark and can change it. Anything the AI is unsure about goes to staff. You can ask for a person to review any mark. |
| CV reading | Reads your CV and suggests values for your profile, such as headline, skills, years of experience, target roles and region. If you ask for ideas to improve your CV, it also suggests up to five changes. | The text of the CV you upload, when you ask us to fill your profile from it or to suggest improvements. For improvements, the kinds of work you are aiming for go with it. Not your name or profile. | Anthropic (Claude), USA. | You check and edit every field before anything is saved. The AI decides nothing about you. The CV is not used to train AI. |
| Skill-check scenario marking | Marks the one written scenario answer in a skill check, first, against the marking guide on How we vet. Score out of 10, with feedback. | The scenario and your written answer. Not your name or profile. | Anthropic (Claude), USA. | Staff review the result before any Vetted badge is awarded. If you do not pass, you can ask for a person to review it, retake after 7 days, and contest the result. |
| Evidence summary for staff | Writes three to five neutral sentences from your evidence list, for the staff member who reviews your vetting. | Your evidence list: certificates, assessments, simulations and skill-check results, with dates. | Anthropic (Claude), USA. | Staff can ignore it and decide differently. It uses the evidence only and no adjectives about you. |
| Job-ad drafting for employers | Drafts an advert from the title and bullet points an employer types. | What the employer typed. No candidate data. | Anthropic (Claude), USA, or the fallback providers in the privacy notice. | The advert is labelled as an AI draft. The employer owns the final text and is responsible for it. Our wording check for discrimination and fees still runs. |
| Enquiry summaries for staff | Summarises an enquiry you send us so staff can reply faster. | The text of your enquiry. | Anthropic (Claude), or OpenRouter when Anthropic cannot answer, USA. | It makes no decision about you. A person replies. |
2. Jobs for you does not use AI
The "Jobs for you" suggestions work by fixed rules. They match your target roles, skills, region and preferences to jobs, and every card says why it was suggested. They are not a decision about you and they never hide a job or stop you applying. If we add an AI model to them, we will update this page first.
3. What AI never does here
- It never ranks or scores candidates for employers. Employers filter the talent pool and sort it by fixed rules they choose. No AI ranking.
- It never rejects a job application and never chooses who an employer hires. Employers decide, and we do not use AI to score your application for them.
- It never works out your gender, ethnicity, age, nationality or other protected characteristics from your name or your CV.
- It never awards a Vetted badge. A named member of staff does.
- Your CV, answers and chats are not used to train AI models, by us or by the providers above.
The one place where an AI mark can affect whether employers can find you is the skill-check scenario. A "not yet" result can stop you being Vetted. So we treat it as an automated decision that matters, and apply the safeguards in the next section.
4. Your safeguards and how to contest a result
These follow the automated decision rules in the UK GDPR as changed by the Data (Use and Access) Act 2025. You can:
- Be told. We tell you before the skill check that AI marks your written answer, and what it looks at.
- Make your case. Add anything you want us to consider when you ask for a review.
- Get a person. Ask for a human review from the result page. A member of staff looks at your answers and the AI's marking, and can change the result.
- Contest it. If you disagree after review, appeal once in writing, then use the complaints procedure. We acknowledge a data protection complaint within 30 days.
Marks from job simulations and module projects have no legal or similarly significant effect, but you can ask a person to review any of them too. To ask, email info@wartens.com with the subject "AI review".
5. Adjustments and fairness checks
You can ask for adjustments to the skill check, such as extra time or assistive technology, at info@wartens.com.
We carry out a data protection impact assessment for CV reading, skill-check marking and the evidence summary, and review it every year. We test the marking for unfair differences between groups where we hold lawful data to do so, including for people who do not write English as a first language and people with dyslexia. The marking guide is about the engineering, not spelling or grammar. If a test shows a problem, we change the marking, not the people.
6. Changes
When we add or change an AI feature, we update this page and the date at the top before the feature goes live. Every notice shown beside an AI feature matches this page. Read how we vet for the standard behind the Vetted badge.
Related pages
Wartens Ltd, trading as EDWartens UK. Version of 28 October 2026. Questions about this page: info@wartens.com. All legal pages.