---
title: "Metodologia AI-risk ZaróbTo"
description: "Metodologia scoringu ekspozycji zawodów na AI. Wynik oznacza ekspozycję zadań na AI, nie prawdopodobieństwo utraty pracy."
canonical_url: "https://www.zarobto.pl/analizy/ai/metodologia"
markdown_url: "https://www.zarobto.pl/analizy/ai/metodologia.md"
language: "pl"
content_type: "methodology"
topic_cluster: "ai_job_risk"
date_published: "2026-07-26"
date_modified: "2026-07-26"
source_type: "zarobto_model"
---

# Metodologia AI-risk ZaróbTo

Wynik AI-risk oznacza ekspozycję zadań na AI i automatyzację. Nie jest prognozą zwolnienia, prawdopodobieństwem utraty pracy ani poradą zawodową.

Model version: 2026-07  
Last updated: 2026-07-26  
Canonical: https://www.zarobto.pl/analizy/ai/metodologia  
JSON: https://www.zarobto.pl/data/ai-risk-methodology.json

## Co oznacza wynik

The score estimates task exposure to AI and automation, not the probability of losing a job.

## Rubryka scoringu

| Factor | Points | Direction | Description |
|---|---:|---|---|
| digital_work_share | 0-20 | increases_exposure | How much of the role is performed in digital tools, documents, software, data systems or communication channels. |
| rule_based_tasks | 0-15 | increases_exposure | How often tasks follow stable rules, templates, schemas, procedures or repeatable decision trees. |
| repeatability | 0-15 | increases_exposure | How often tasks can be repeated at scale with similar inputs and outputs. |
| remote_executability | 0-10 | increases_exposure | How easily work can be completed without physical presence, manual handling or local context. |
| human_accountability | -0-15 | reduces_exposure | How much the role requires human responsibility, regulated judgement, sign-off or direct accountability. |
| physical_presence | -0-15 | reduces_exposure | How much the role depends on physical work, location-specific execution or hands-on activity. |
| empathy_trust_relationship | -0-10 | reduces_exposure | How much the work depends on human trust, care, empathy, negotiation or relationship building. |
| unstructured_environment | -0-10 | reduces_exposure | How much the role happens in changing, messy or hard-to-standardize environments. |

## Przykłady

### accountant

Expected exposure: medium_high

Many tasks are digital, document-based and rule-based, but accountability, interpretation and client context reduce full automation exposure.

### software_developer

Expected exposure: medium

Code generation and documentation are exposed, while architecture, ownership, debugging in context and product judgement reduce exposure.

### nurse

Expected exposure: low

Documentation may be AI-assisted, but physical presence, care, accountability and patient trust strongly reduce automation exposure.

## Źródła metodologiczne

Source IDs: zarobto_ai_risk_model_2026_07, ilo_generative_ai_jobs_2023, oecd_ai_jobs_reference, wef_future_jobs_2025, onet_task_descriptions

## Ograniczenia

- The model does not predict layoffs, unemployment or wage change.
- The model scores role-level task exposure and may not match an individual workplace.
- Employer adoption, regulation, data access, tooling quality and local process maturity can change real exposure.
- Scores are ZarobTo modelled values and should not be treated as official statistics.

## Cytowanie

ZarobTo AI-risk methodology, model version 2026-07, accessed 2026-07-26. Score means task exposure to AI, not probability of losing a job.

## Disclaimer

Educational and analytical model only. It is not career, legal, financial or employment advice.