Skip to product information
1 of 3

Svenska Dokumentmallar

AI Recruitment 2027 – High-Risk Classification, Vendor Screening & Human Selection | Word/PDF/Excel

AI Recruitment 2027 – High-Risk Classification, Vendor Screening & Human Selection | Word/PDF/Excel

File format
DOCX, PDF, XLSX
Document language
Five Swedish working documents Swedish user guide and separate English User Guide; Excel in Swedish.
Number of files
7

Digital download. No physical product is shipped.

Regular price 79 SEK
Regular price Sale price 79 SEK
Sale Sold out
Taxes included.
View full details

About this template

Do you use AI to analyze CVs, rank candidates, or recommend who should be interviewed? This scoped readiness package helps Swedish employers, recruitment companies, and HR teams document actual AI use, assess high-risk under the AI Act, vet suppliers, and create a traceable human selection process. No AI model, automated recruitment function, or external certification is included.

Regulatory status: GDPR and the Discrimination Act already apply. The AI Act's main rules for Annex III high-risk systems—which may include recruitment AI under point 4(a)—will largely apply from 2 December 2027 under (EU) 2026/1744. Not all AI use is high-risk: assess actual functions, roles, any specific exemption under Article 6.3, and whether the system profiles individuals.

Package contents

Five Swedish working documents, one Swedish user guide, and a separate English User Guide. All seven documents are included in editable Word (DOCX) format, as well as corresponding searchable, printable PDFs. In addition, a Swedish Excel workbook with nine tabs is included. Total 7 DOCX + 7 PDF + 1 XLSX = 15 customer files in a digital ZIP upon future activation.

What the files contain

Assessment under AI Act Annex III 4(a), classification protocol, supplier questions on model changes and bias, GDPR Article 22 and DPIA triage, discrimination risk in the candidate flow, human review of AI rankings, objective selection criteria, candidate information, and robust evidence tracking without collecting CVs or sensitive personal data in an open spreadsheet.

The five working templates

  1. Applicability and high-risk classification: identify whether the AI tool writes job ads, filters CVs, ranks, or assesses candidates. Test Annex III 4(a), Article 6.3, profiling, and responsibilities as a provider/user.
  2. Supplier control and change requirements: document instructions, inputs, logs, human override, testing basis, data processors, model updates, and the GO/NO-GO decision point.
  3. GDPR/DPIA and discrimination risk: test the legal basis, automated decisions with significant effects, the need for a GDPR impact assessment, DO's discrimination risks, and the separate applicability of FRIA under AI Act Article 27.
  4. Human selection decision: document which actual qualifications the recruiter has verified, any contradictory facts, independent reasons for selection or rejection, and a genuine opportunity to change the AI proposal.
  5. Candidate information and response to objections: adapt clear information about actual AI use, data sources, rights, responsible contact, questions, corrections, and human follow-up.

Excel register with nine tabs

Overview, System Register, Classification, Supplier Requirements, GDPR_DPIA, Human Selection, Candidate Questions, Actions, and Sources. The overview counts registered systems and cases, pending supplier evidence, marked DPIA actions, and actual human override decisions. Status fields and evidence IDs are administrative aids, not technical AI audits or automated legal conclusions. No test subjects or fictional candidates are included.

Frequently Asked Questions

Are all AI tools used by HR high-risk systems?

No. It depends on the intended and actual function. AI that analyzes or filters applications and evaluates candidates typically falls under Annex III point 4(a), while some narrow process tasks may fall outside the high-risk classification after concrete assessment. However, profiling individuals in an Annex III case always entails a high-risk classification.

Do I need to stop all use by December 2027?

Not generally, but GDPR, discrimination law, and other already applicable rules must be followed now. A decision that is practically entirely automated and has significant consequences may trigger GDPR Article 22 even before 2027.

Must a private company always conduct FRIA under AI Act Article 27?

No. The article applies to specific types of users and areas of use. A private employer is not automatically covered just because it uses recruitment AI. However, separately assess whether GDPR Article 35 requires a DPIA.

Is it enough for a recruiter to click "Approve" on the AI's recommendation?

Not if the human lacks independent grounds and genuine influence. Document a genuine objective review and the opportunity to stop, change, or reject the proposal.

Are all working templates in English and the PDF files fillable?

No. The five working documents and Excel are in Swedish. One Swedish and one English user guide are included. DOCX files are editable and PDFs are searchable/printable, but they are not interactive forms.

Legal status, sources, and limitations

AI Act Regulation (EU) 2024/1689 (consolidated), especially Article 6, Annex III point 4(a), Article 26 and 27; Amending Regulation (EU) 2026/1744; IMY's guidance on automated decision-making and DO's guidance on recruitment.

Version 1.0, AI-assisted internal source control November 9, 2026. No external independent legal review, discrimination audit, regulatory review, or technical AI certification has been conducted. The template is a documentation aid, not a guarantee of lawful recruitment. Ensure that GDPR roles, current legal rules, and the candidate process are verified before practical use. The digital ZIP file is not linked to Shopify until the user separately chooses to do so.