Our current age is compared to the gilded age of the late 19th century, in which national-level industry emerged. A “progressive era” followed that age, which created political mechanisms that better supported the new national industry landscape. I speculate that a new progressive age, based on the dominant industry of our time—IT/artificial intelligence—may arise to reform the political landscape to better support the current possibilities. Possible reforms will be discussed and likely difficulties in adopting them will be considered.
Evaluative artificial intelligence (AI) systems are widely used to assess, rank, and score individuals in high-stakes contexts, raising significant human-rights concerns. This exploratory study examines perceived governance challenges using a practice-oriented approach grounded in international AI governance frameworks. Applying importance–performance analysis to expert perceptions, the results reveal differences between the perceived importance and current implementation of eight ethical governance dimensions. Transparency emerges as the only relatively high-importance/low-performance priority, although its position is close to the sample-based quadrant boundaries. The findings suggest perceived implementation shortfalls across several dimensions and provide practical guidance for strengthening accountability, oversight, and human-centered governance mechanisms.
This article defines a continuum of human–machine interaction—from human-in-the-loop (HITL), through human-on-the-loop (HOTL), to human-out-of-the-loop (HOOTL)—and argues that competitive advantage increasingly favors moving decisions along that continuum toward full machine autonomy. Executives who insist on human oversight in domains where machines now reliably outperform their employees are not exercising prudent caution; they’re making costly leadership mistakes. The article outlines three steps executives can take to exploit this shift: First, close the gap between artificial intelligence (AI) "power users" and "light users" within their organizations; second, systematically and continuously identify which decisions machines already make better than humans, tracking that boundary forward across multiple time horizons; and third, prototype HOOTL opportunities the moment they cross that threshold, rather than waiting for industry-wide consensus on when AI becomes acceptable. Together, these steps describe a new executive core competency.
Shadow artificial intelligence (AI), defined as the unsanctioned use of generative AI within organizational environments, has become a structural challenge for modern enterprises. While organizations increasingly deploy governed AI solutions, employees continue to leak proprietary data through public chatbot interfaces, creating irreversible exposure through model absorption, the process by which training data become permanently embedded in model parameters. This article moves beyond traditional binary blocking approaches, proposing an adaptive governance matrix that balances data sensitivity against business innovation value. We introduce the concept of deep shadow AI, where prohibition drives adoption underground, and present a comprehensive AI Gateway architecture for auditable large language model integration. The framework provides IT leaders with actionable checklists aligned with International Organization for Standardization/International Electrotechnical Commission 42001 standards and the European Union AI Act regulatory requirements.
