Digital HR tools — from AI to VR — can raise engagement and on‑the‑job performance by enhancing autonomy, competence and relatedness, but gains are inconsistent and hinge on privacy safeguards, explainability and workers' digital skills.

Main Finding

The PRISMA-guided systematic review finds that technology-driven positive-psychology interventions (notably AI, VR/AR, wearables, and chatbots) can enhance employee well‑being, engagement, and organizational performance, but effects are non-linear and context-dependent. The authors propose the Technology-Enhanced Well‑Being Framework (TEWF) — integrating affordance theory, Self‑Determination Theory (SDT), and PERMA — to explain how digital HR affordances shape autonomy, competence, and relatedness and thereby affect motivation and productivity. Adoption and outcomes are strongly moderated by privacy, explainability, digital literacy, and regulation; ethical risks (AI bias, data privacy) can undermine benefits.

Key Points

• Scope and evidence base
  • Systematic review of studies published 2021–2025 across Asia, Europe, and the U.S.
  • Abstract reports 47 peer‑reviewed studies; PRISMA flow in the paper indicates 49 studies included (authors note this inconsistency).
• Theoretical contribution
  • Technology‑Enhanced Well‑Being Framework (TEWF): links digital affordances → SDT needs (autonomy, competence, relatedness) → PERMA well‑being outcomes → motivation/performance.
  • Uses affordance theory to explain how HR tech features enable or constrain psychological processes.
• Types of interventions
  • AI-enabled recruitment and talent management, performance analytics, chatbots for support, VR/AR for learning and mental‑health interventions, wearables for physiological/EMA monitoring.
  • Taxonomy developed of AI‑enabled HR interventions (thematic classification rather than a single prescriptive list).
• Empirical patterns and moderators
  • Positive effects on engagement, resilience, emotional intelligence, and self‑efficacy are reported, but relationships are often non‑linear and contingent on context.
  • Key moderators: data privacy protections, algorithmic explainability, employee digital literacy, and regulatory environment.
• Methodological observations
  • Common methods include SEM, PLS‑SEM, meta‑analysis, and Ecological Momentary Assessment (EMA).
  • Many primary studies are cross‑sectional; longitudinal and multicultural work is limited.
• Ethical and governance concerns
  • Algorithmic bias, fairness, and data privacy are recurring risks; the authors advocate ethics policies, digital‑literacy programs, and technology audits.
• Practical output
  • A decision framework for HR leaders to evaluate and adopt technology‑driven positive‑psychology strategies.

Data & Methods

• Review approach: Systematic Literature Review following PRISMA guidelines, framed with CIMO (Context, Intervention, Mechanism, Outcome).
• Search and selection
  • Primary database: Scopus (chosen for indexing quality); supplementary sources used.
  • Keywords: “HR Technology,” “PERMA,” “AI in HR.”
  • Initial Scopus hits: 302 records. After exclusions (date range 2021–2025, indexing quality, duplicates, abstracts), 185 remained for screening.
  • Screening & retrieval: authors report records screened = 185; many records not retrievable (74). Combined Scopus + other sources yielded 49 studies assessed for eligibility and included (paper text contains a small discrepancy between 47 vs. 49 included studies).
  • Tools: collaborative platform Watase.web.id used to coordinate searches and management.
• Analytic approach
  • Thematic synthesis to build taxonomy and develop TEWF.
  • Review integrates empirical results (quantitative and qualitative) and examines mechanisms via SDT and PERMA lenses.
• Limitations noted by authors
  • Retrieval gaps (many studies not obtained), reliance on Scopus limits some coverage, prevalence of cross‑sectional designs, limited geographic/demographic diversity in primary studies.

Implications for AI Economics

• Productivity and returns to AI investment
  • Technology that increases employee well‑being and engagement can raise labor productivity and reduce turnover—implying positive returns to AI/HR tech investment. However, gains are conditional on complementary investments (digital literacy, privacy protections, explainability).
  • Non‑linear effects mean marginal returns vary by context and may exhibit thresholds: below certain levels of trust or digital skill, AI can harm productivity (via stress, mistrust), while above thresholds it generates outsized gains.
• Labor demand, skill composition, and human capital
  • Emphasis on digital literacy and explainability highlights a shift in human capital demands: employers will value employees capable of working with AI tools and interpreting algorithmic outputs. Training and continuous learning become economic inputs.
  • Strengths‑based and positive‑psychology uses of tech (e.g., role redesign, neurodiversity inclusion) can alter task allocation and complementarities between humans and machines.
• Adoption, diffusion, and regulation
  • Privacy and fairness regulation materially affect adoption costs and technology value. Stricter data/algorithm regulations can raise compliance costs but may increase trust and effective uptake, altering adoption trajectories across firms and sectors.
  • Economic models of AI diffusion should include institutional and regulatory moderators, as well as firm‑level governance capacity.
• Market failures and externalities
  • Algorithmic bias and privacy harms create distributional risks and potential market failures (discrimination, reputational damage). There is a role for audits, standards, and possibly public subsidies to internalize these externalities (e.g., support for audits, training).
• Measurement and evaluation
  • The TEWF provides a mechanism‑based structure that economists can use to model how HR tech affects utility (well‑being) and productivity. Incorporating PERMA‑type well‑being metrics and EMA data into empirical economic analyses allows more accurate welfare accounting beyond output measures.
  • Cost‑benefit and ROI analyses of HR AI should include direct productivity effects, turnover and recruitment costs, training costs, privacy/compliance costs, and expected loss from biased or opaque systems.
• Policy recommendations for economic actors
  • Firms: invest in complementary human capital (digital literacy), explainable systems, and privacy safeguards to capture productivity gains.
  • Policymakers: encourage transparency standards, support longitudinal research and data access for independent evaluation, and consider targeted incentives (or mandates) for ethical audits to reduce negative externalities.
• Research opportunities for AI economics
  • Quantify the marginal productivity returns to wellbeing‑oriented HR tech and their heterogeneity across firm size, sector, and workforce skill composition.
  • Model threshold/non‑linear adoption dynamics and regulation impacts on welfare and inequality.
  • Evaluate long‑run labor‑market effects of using AI for selection/promotion with attention to bias and sorting.

If you want, I can (a) produce a concise one‑page executive summary in Portuguese, (b) extract the thematic taxonomy of AI‑enabled HR interventions into a table, or (c) outline a simple economic model that formalizes TEWF mechanisms for empirical testing. Which would you prefer?