Surfing the AI waves: the historical evolution of artificial intelligence in management and organizational studies and practices

Matteo Cristofaro and Pier Luigi Giardino ask why advances in AI, management scholarship and organizational adoption repeatedly move at different speeds. Their historical literature synthesis proposes five overlapping waves to explain these mismatches. It first appeared online on March 19, 2025; the Journal of Management History issue is dated June 23, 2026.^[1]

Method and framework

The authors searched Scopus and Web of Science for English-language peer-reviewed work in journals ranked 4 or 4* in the 2024 Academic Journal Guide. Their search combined current terms such as machine learning and big data with earlier terms such as cybernetics. They describe the result as a targeted historical synthesis rather than an exhaustive review. Economic long-wave theories, technological paradigm shifts and hype cycles supply interpretive models.^[1]

The five phases are symbolic AI and theoretical foundations; the AI winter and narrow applications; the machine-learning renaissance; big data and deep learning; and an emerging human–AI collaboration and responsible-AI phase. Within each, the paper compares technical developments, scholarly contributions, applications and barriers. The fifth phase includes a proposed research agenda, rather than a completed historical period.^[1]

The central argument is that better algorithms alone do not ensure useful adoption: organizations also need infrastructure, skills, trust and suitable decision processes. Table 2 connects unequal readiness to recommendations for investment in AI literacy and infrastructure. Table 1 turns the collaboration theme into questions about which tasks to delegate, how hybrid decisions affect errors and how governance can address opaque or biased outputs. These are research questions and recommendations, not demonstrated effects of the framework.^[1]

Limits

The authors acknowledge that broad waves can obscure incremental changes and differences between sectors and regions. Selecting journals by their current rankings can exclude interdisciplinary work and journals that mattered more in earlier periods. The synthesis does not establish a measured causal relationship between institutional readiness and adoption, or validate a forecast of the next wave.^[1]

Its chronology also needs care: section 3.1 places the coining of artificial intelligence at the 1956 Dartmouth workshop, although the term already appears in the title and opening of the proposal dated August 31, 1955.^[1][2]

The paper offers an organizational perspective on Social Alignment: adoption involves choices about work, governance and human participation as well as technical capability. Its theory–practice mismatch concerns institutions and scholarship, distinct from the learned-objective problem described by Inner Alignment. The authors’ practical conclusion is to treat adoption as continuing organizational learning, with human expertise and accountability built into the process.^[1]