AI is moving from operational tools into the strategic core of firms—boosting innovation where governance and skills exist but concentrating power with data controllers and amplifying bias and workplace surveillance. Latin American companies face uneven adoption due to skill, data and regulatory gaps, so responsible integration demands human oversight, algorithmic transparency and training aligned to regional inequalities.

Main Finding

The paper (Sevilla Godínez, 2026) concludes that AI adoption is reshaping firms’ managerial rationality, concentrating power with data controllers, shifting professional judgment toward opaque models, and generating ethical, labor, and governance risks. Responsible adoption in Latin America requires organizational governance (human oversight, algorithmic transparency) and targeted training policies that account for regional inequalities.

Key Points

• Scope and corpus: Narrative review of 54 sources with an analytical corpus of 26 peer‑reviewed articles (2022–2026) drawn from Scopus, Web of Science, Google Scholar, SciELO, Redalyc and Dialnet.
• Strategic move inward: AI has moved from operational tasks to the firm’s strategic core (decision-making, product innovation, customer experience).
• Firm performance channel: Evidence links AI investment to higher sales, employment and market valuation, with product innovation a primary channel (e.g., Babina et al., 2024).
• Heterogeneous productivity effects: Field evidence shows productivity gains (e.g., conversational assistants raising productivity ~15% in one study), but benefits concentrate among certain worker groups (less experienced or those with complementary skills).
• Decision-making limits: Human–AI combinations do not always outperform the best component; selective adherence to algorithmic advice and difficulties delegating cognitive tasks are common (meta‑analysis and experiments cited).
• Opacity and bias: Black‑box models produce explainability problems and algorithmic injustices that can induce discriminatory organizational decisions without commensurate guilt or accountability.
• Labor and well‑being: AI changes job content more than it eliminates jobs outright—reshaping roles, intensifying surveillance, eroding boundaries (employee vs. freelancer), and increasing psychosocial costs (isolation, insomnia).
• Governance shortfall: Firms often lack formal AI oversight policies; declared ethical principles are insufficient without institutional design that enables appeal, accountability and stakeholder‑tailored transparency.
• Latin American context: Adoption is fragmented and unequal—limited by digital infrastructure, scarce skilled personnel, weak regulation and concentration in particular functions (e.g., marketing, finance, services).
• Management algorítmico risks: Algorithms create a “gray zone” enabling regulatory evasion, de facto power shifts toward data owners, and novel compliance challenges.

Data & Methods

• Design: Qualitative narrative review (appropriate for rapidly evolving, multidisciplinary literatures).
• Search strategy: Bilingual keyword searches (Spanish/English) across major international and regional databases (Scopus, WoS, Google Scholar, SciELO, Redalyc, Dialnet).
• Inclusion criteria: Peer‑reviewed articles published 2022–2026 addressing organizational transformation, ethics, governance or labor impacts of AI.
• Screening & selection: From >150 initial records, title/abstract screening plus full‑text review produced a final analytic corpus of 26 studies (reported in a detailed Table 1), complemented by a broader set of 54 sources supporting background.
• Evidence types in corpus: Systematic reviews, empirical quantitative and qualitative studies, field experiments, conceptual essays and sectoral case studies covering global and Latin American contexts.
• Analytical approach: Thematic synthesis across five categories—structure & decisions, opacity & bias, work reconfiguration, governance frameworks, and Latin American adoption conditions.
• Limitations: Narrative (not meta‑analytic), selective inclusion emphasizing recent literature (2022–2026), and heterogeneous evidence quality across case studies and experiments. Findings are interpretive rather than statistically generalizable.

Implications for AI Economics

• Firm‑level productivity and value:
  • AI can raise firm sales, market valuation and employment via innovation and service redesign, but gains are uneven across firms and worker types. Empirical work should isolate causal channels (R&D vs. adoption intensity vs. complementary human capital).
• Labor markets and distribution:
  • AI reallocates tasks within jobs and across occupations; complementarities matter. In Latin America, weak digital skills and infrastructure may limit complementarities, increasing inequality. Economists should measure task‑level exposure, wage‑skill complementarities, and distributional consequences regionally.
• Market power and data concentration:
  • Redistribution of decision power toward data controllers implies concentration externalities and potential market power. Antitrust and industrial organization research should assess data‑driven market frictions, entry barriers, and the competitive effects of platformized algorithmic management.
• Measurement and evaluation:
  • Standard productivity metrics may understate or misattribute AI gains (e.g., quality of service, fast product cycles). Microdata—firm panels, matched employer–employee datasets, and platform logs—are crucial for accurate impact estimates.
• Ethical and regulatory economics:
  • Algorithmic bias, opaque decision rules, and surveillance have economic effects (e.g., discriminatory hiring reduces aggregate productivity via misallocation). Policy design should incorporate enforceable transparency, appeal mechanisms, and governance that internalizes social costs.
• Policy priorities for emerging economies:
  • Invest in data infrastructure, skills training targeted to complementarities with AI, and institutional capacity for algorithmic oversight. Subsidies or public procurement to spur diffusion should be paired with governance requirements (explainability, audits).
• Research agenda:
  • Causal, sector‑specific studies on productivity, employment and welfare impacts in Latin America.
  • Analyses of how firm governance (board oversight, data governance) moderates AI returns.
  • Models of market structure with data‑driven entry barriers and algorithmic coordination.
  • Evaluation of policy instruments (training programs, transparency mandates, data portability, antitrust interventions) under resource constraints typical of the region.

Reference: Sevilla Godínez, H. (2026). "Inteligencia artificial en la empresa contemporánea: implicaciones filosóficas, éticas y organizacionales." Enfoques. Revista de Investigación en Ciencias de la Administración, 10(39), pp. 1–18. DOI: 10.33996/revistaenfoques.v10i39.252.