Can AI Make Our Supply Chains Fairer?
From forced labour in solar panel supply chains to exploitative conditions in fast-fashion, conventional auditing has failed workers. As new legislation like the EU's Corporate Sustainability Due Diligence Directive demands deep supply chain visibility, companies are turning to AI. This page presents an interactive summary of a systematic review that asks: does AI protect workers, or create new vulnerabilities?
Figure 1: 94% of reviewed articles were published between 2022 and 2025, reflecting the topic's rapid emergence.
Figure 2: "General AI" dominates (41 papers), but ML, CV, and NLP are gaining specificity.
The Duality of AI's Impact on Labour Principles
The research reveals a fundamental duality: 54% of the literature documents AI as a tool for worker protection, while 40% highlights it as a catalyst for new forms of precarity and control.
AI as a Tool for Protection
AI functions as a powerful analytical lens, identifying patterns invisible to conventional auditing. Industry 5.0 integrates cognitive computing with automated robotics to redesign hazardous tasks and boost worker wellbeing.
- Proactive Safety: Predictive analytics forecast supplier adherence and identify hazards before accidents happen.
- Detecting Forced Labour: ML-based screening pierces structural opacity to flag modern slavery risks.
- Transparency: NLP and LLMs aggregate unstructured data into actionable ESG insights.
AI as a Catalyst for Precarity
AI-driven automation and algorithmic management threaten the foundations of decent work. The AI development supply chain itself relies heavily on unregulated crowdwork platforms with low, volatile wages.
- Job Displacement: 27% of current jobs are highly susceptible to automation (OECD); generative AI outperforms crowdworkers at a fraction of the cost.
- Wage Suppression: Deskilling expands the low-skill labour pool, diminishes bargaining power, and drives wage stagnation.
- Hyper-Surveillance: Workers labelled "idle" for biological needs, coerced into unpaid overtime by algorithmic performance targets.
AI's Impact on the 9 ETI Base Code Principles
The study maps AI's impact against the nine ETI Base Code principles across 46 papers that explicitly connect AI to discrete social dimensions. "Safe and Hygienic Work" dominates (32 papers), while intangible rights like "Freedom of Association" and "Working Hours" remain under-researched.
Click on a bar to see the specific findings for that principle.
An Integrated TOE-DOI Framework for AI Adoption
The study synthesises the Technology-Organisation-Environment (TOE) framework with Diffusion of Innovations (DOI) theory into a novel three-stage lifecycle model, yielding six formal propositions. Use the controls below to explore each stage.
Epistemic Readiness
Decision & Implementation
Ethical Governance
Future Research Directions
The study identifies four research themes with eight proposed research questions, shaped by the most pressing gaps in the literature.
1. Deepen Empirical Specificity
67% of the literature focuses on observable compliance while only 10% addresses intangible structural rights. Future work should test how specific AI tools (ML, NLP, CV) impact intangible labour rights and examine how task simplification simultaneously improves safety while degrading worker autonomy.
2. Navigate the Digital Sustainability Divide
High costs and poor data quality create a severe divide between lead firms and SMEs. Research should investigate how this divide structurally impedes ethical AI adoption, and under what conditions ethical AI policies inadvertently produce algorithmic greenwashing.
3. Broaden Theoretical Perspectives on GVCs
21% of papers rely on the Resource-Based View, framing AI as a competitive tool while obscuring power asymmetries. Alternative lenses like Labour Process Theory should examine how supply chains perpetuate "digital colonialism" and hold lead firms accountable for precarious data workers.
4. Advance Post-Adoption Trustworthy AI
AI models suffer from concept drift and algorithmic opacity. Research should explore how Explainable AI (XAI) and decentralised technologies can mitigate drift and build trust, and what hybrid governance models ensure continuous ethical calibration after deployment.