Explainable AI (XAI) in Digital Marketing: A Systematic Review

An independent systematic literature review synthesizing recent research on the application and impact of explainable AI (XAI) in digital marketing, with a stakeholder-oriented analytical framework.

Research Snapshot

Research Type:
Systematic Literature Review (SLR)

Domain:
Explainable AI · Digital Marketing · Human-Centered AI

Timeframe Covered:
2023–2025

Corpus:
268 articles retrieved

Methods:
Thematic synthesis and research question synthesis

Status:
Working paper (not yet peer-reviewed)

Research Questions

• RQ1: How has XAI been applied in digital marketing systems to enhance transparency and interpretability of machine learning models?

• RQ2: What impacts of XAI adoption in digital marketing have been reported, particularly regarding trust, decision-making, fairness, and accountability?

Methodology

Articles were retrieved from Scopus using a predefined search string applied to titles, abstracts, and keywords. After duplicate removal and multi-stage screening based on inclusion and exclusion criteria, the final corpus was analyzed using thematic synthesis and research question–driven integration.

A structured appraisal framework was used to evaluate methodological rigor, stakeholder focus, explainability techniques, and reported impacts.

Key Findings (bullet points, not prose)

Key Findings:

• XAI in digital marketing is predominantly implemented as a post-hoc interpretability layer on black-box models (e.g., SHAP, LIME).

• Measurable impacts of XAI are predominantly reported for internal stakeholders through performance and interpretability metrics, while end-user benefits remain largely unquantified.

• Reported benefits focus on decision efficiency and performance optimization, while empirical evidence on trust, fairness, and accountability remains limited.

• Ethical considerations are frequently acknowledged but rarely operationalized or evaluated.

Contributions

• Identifies structural gaps between explainability implementation and human-centered evaluation.

• Proposes future research directions bridging XAI, trust calibration, and ethical marketing systems.

Future Work

• Design evaluation frameworks/algorithms that quantify customer trust and engagement to evaluate impact of XAI.