Generative AI is reshaping how firms are organised: literature mapping shows GenAI creating an 'algorithmic workplace' where decision-making becomes more decentralized and managers shift from micro-managers to curators and coordinators. The emerging discourse links technical constructs (LLMs, GANs) to managerial concepts—implying potential productivity gains, changing skill demands, and new governance challenges, but stopping short of causal evidence.

Main Finding

Generative AI (GenAI) is rapidly reshaping the architecture of the workplace and managerial agency. The paper’s bibliometric–conceptual review of 212 peer‑reviewed studies shows a fast‑growing, interdisciplinary field in which technical advances (e.g., LLMs, GANs) intersect with managerial concerns (autonomy, coordination, decision‑making). The authors propose an “algorithmic workplace” framework characterized by hybrid human–algorithmic agency, decentralised decision‑making, and a shift from command‑and‑control to guide‑and‑collaborate managerial paradigms.

Key Points

• Rapid expansion: the literature on GenAI and workplace transformation exploded during 2023–2025 (annual growth ≈115%), moving from early descriptive studies to emergent conceptual clusters.
• Intellectual structure: six coherent conceptual clusters were identified (technical constructs, human–AI teaming, algorithmic authority, ethics/transparency, capability building, open innovation dynamics) and four meta‑themes synthesized: (1) Algorithmic Authority & Control, (2) Human–AI Teaming & Coordination, (3) Ethics/Transparency/Accountability, (4) Capability Building & Learning.
• Managerial implications: GenAI reduces the marginal cost of many cognitive tasks, expands employee autonomy, and can weaken informational asymmetries that historically supported managerial layers—promoting flatter, more modular organizational forms.
• Socio‑technical role: GenAI functions as an intermediary in decision support—altering epistemic work, redistributing cognitive labour, and creating hybrid decision processes where humans and models co‑produce outcomes.
• Distributional effects: benefits are uneven—organisations and workers with higher digital readiness and absorptive capacity capture greater gains; risks include accountability gaps, opacity, and erosion of relational leadership.
• Field dynamics: research is interdisciplinary and concentrated in a few core journals (e.g., Expert Systems with Applications, Sustainability). A small group of authors plays an outsized role in structuring discourse; keyword and Sankey analyses show lexical convergence across subfields.
• Normative and governance concerns: calls for transparency, human oversight, explainability, and regulatory attention are central and urgent given that early deployment decisions tend to harden into organizational routines.

Data & Methods

• Corpus: 212 peer‑reviewed articles identified through systematic searches of Scopus and Web of Science (search window and exact inclusion dates reported in paper; screening resulted in 212 included studies).
• Search & screening: PRISMA‑informed protocol; inter‑rater reliability κ = 0.87 reported for inclusion decisions.
• Tools: Bibliometrix (R) for performance metrics; VOSviewer for network and science mapping (keyword co‑occurrence, co‑citation, country co‑authorship).
• Parameters: full counting, minimum threshold of five occurrences for network clustering (chosen to balance noise vs. thematic diversity); temporal overlays to trace evolution.
• Analysis approach: mixed bibliometric mapping followed by theory‑driven interpretive synthesis linking clusters to organizational, innovation, and socio‑technical theories.
• Reproducibility: authors report archiving scripts, datasets and parameter settings; Data & Code Availability details in appendices.

Implications for AI Economics

• Task allocation and labor composition: GenAI lowers costs of many cognitive tasks (writing, coding, summarising), accelerating task reallocation within firms. Expect shifts in demand toward complementary skills (supervision, model evaluation, domain judgement) and potential displacement of intermediate cognitive tasks—implications for wage structure and skill‑biased technological change.
• Productivity and measurement: GenAI’s impact on productivity will depend on firms’ ability to integrate models into decision workflows and governance. Economics research should develop precise productivity metrics that separate automation gains from coordination and quality effects.
• Organization of firms and coordination costs: Algorithmic coordination can lower the information processing function of hierarchical managers, which changes optimal firm size and internal boundaries. Models of the firm should incorporate algorithmic intermediaries when analysing make vs buy, vertical integration, and decentralisation.
• Returns to digital readiness and inequality: Gains are likely concentrated among firms (and workers) with higher absorptive capacity, amplifying within‑ and between‑firm inequality. Empirical work is needed to quantify heterogeneity in returns and spillovers across sectors and regions.
• Value capture and market structure: As GenAI is embedded in strategic processes, questions arise about who captures the surplus—platform providers, firms deploying AI, or workers. Industrial organization models should consider how GenAI affects market power, switching costs, and complementarities with proprietary data.
• Governance, regulation and externalities: Policy has to address transparency, liability, and human oversight. Economists should model regulatory impacts (e.g., disclosure rules, liability regimes) on adoption, innovation incentives, and social welfare.
• Research priorities for AI economics drawn from the paper:
  • Causal, field‑level studies measuring effects of GenAI on employment composition, wages, productivity, and firm performance.
  • Structural models of firm organization incorporating algorithmic agents and hybrid decision rules.
  • Distributional analyses across worker skill groups and firm types to assess inequality and labor market adjustment costs.
  • Welfare analysis of governance interventions (transparency mandates, audit requirements) balancing innovation incentives and accountability.
  • Empirical work on complementarities between GenAI and organizational practices (training, process redesign) that determine realized gains.
• Limitations to bear in mind: the evidence base is early and rapidly evolving; bibliometric indicators reflect scholarly attention more than causal impacts. Economists should complement bibliometric insights with longitudinal microdata, randomized trials, and quasi‑experimental designs.

If you’d like, I can: - Extract the six conceptual clusters and their representative keywords/authors from the paper’s visualisations. - Draft empirical research designs (tests, data sources, identification strategies) to study GenAI effects on wages, employment and firm productivity.