The 5 forces creating urgency around ethical AI adoption

Jeffrey N. Saviano

Oct 28, 2026 5 minute read

In his new book, “Boundaries of Tolerance,” MIT Sloan’s Jeffrey Saviano says that AI’s novel risks and “ethical infirmities” have made governance and strategy a leadership necessity.

Artificial intelligence is fueling innovation across industries, but it’s also introducing unanticipated risks and consequences for organizations.

In this landscape, traditional enterprise risk management frameworks are no longer sufficient, and AI governance becomes a leadership question — one that “reaches the boardroom and senior executive teams,” according to Jeffrey Saviano, an MIT Sloan School of Management senior lecturer who spent three decades at EY.

In an excerpt from his new book, “Boundaries of Tolerance: A Foundational Model for Ethical AI Strategy and Governance,” Saviano examines five forces that create urgency for adopting ethical AI solutions and practices.

We are living in an era of unprecedented opportunity and transformation as AI continues to advance at a remarkable pace. AI is fueling innovation across industries, creating new products and reinventing business models almost overnight. However, alongside these benefits, significant concerns continue to emerge about the institutional risks and ethical consequences emanating from AI systems. Leaders across both the public and private sectors are awakening to the complexity AI introduces. As new strategies and applications for AI emerge, so too do novel risks and ethical infirmities that demand careful consideration.

The consequences of these governance gaps are real. Behind every algorithmic failure is a person or broader community that bears the cost. Consider the patient whose medical diagnosis is compromised because an AI model was trained on incomplete or biased clinical data, or the job candidate rejected by an automated hiring system informed by archaic employment patterns. Even small system-input errors can cascade into outcomes that reinforce inequality and erode stakeholder trust.

The Human Consequences of AI

For all the attention given to the technical wonders of AI, its impact remains profoundly human. Every line of code carries an assumption about fairness, and every model release carries a decision about who wins (benefits from the AI system) and who loses (is harmed by the system).

This is why the question of AI governance cannot be confined to technologists or regulators alone. It is a leadership question, one that reaches the boardroom and senior executive teams, because the ethical boundaries established there shape how these systems touch lives far beyond the enterprise. Throughout [my] book, I use the term “ethical AI,” while others often use “responsible AI” to describe similar efforts. The distinction is not critical, but I prefer ethical AI because it emphasizes the role of leadership judgment in how these systems are designed and deployed. Ethical AI strategy has become a leadership necessity. AI systems are moving into core business operations across nearly every industry.

I have observed five forces driving companies to adopt an ethical AI strategy.

First, competitive pressure has made [the] speed of AI adoption feel unavoidable. AI is now embedded in core products and internal operations in many organizations. Companies feel intense pressure to deploy AI quickly to protect market position and signal innovation. In many organizations, the fear of falling behind competitors outweighs concern about deploying systems that are not fully understood. As a result, AI systems are entering production environments before organizational leadership has clarified how they should be governed. Early choices about system design and deployment are being made under time pressure. Those choices can result in long-term exposure for the enterprise.

Second, AI has changed how risk behaves within the modern enterprise. With traditional systems, small problems typically remain small. With AI, minor issues can escalate rapidly once systems are deployed throughout the organization. A biased outcome affecting a limited population can trigger public backlash within hours. A model that performs acceptably in testing can behave differently in real-world use. These dynamics undermine familiar assumptions about materiality and proportionality. Many companies are sensing that existing oversight models do not match the behavior of modern AI systems. This creates a growing awareness that traditional enterprise risk management frameworks are no longer sufficient.

Third, expectations from investors and markets are evolving. Questions about AI oversight are beginning to resemble earlier scrutiny of cybersecurity and associated internal controls. Investors are asking who oversees AI risk and how failures are addressed. These questions often arise after incidents result from their deployed systems. Professionals are expected to explain outcomes without having been given clear standards for harm prevention. This gap increases pressure on leadership to define its own approach. Ethical AI governance is becoming a defining element of effective leadership.

Fourth, pressure is also coming from inside the organization. Employees across functions are already using AI tools in their daily work. They are encountering ethical questions in real time. They ask what types of data should fuel their AI models. They ask whether specific AI systems should be deployed. They ask who is responsible when harm occurs. In many organizations, these questions are escalated upward because decision authority is unclear. When leadership has not defined appropriate boundaries, employees either stop raising concerns or make decisions on their own. Both outcomes weaken organizational control.

[Fifth], moral expectations about AI use are rising. Public conversations about AI increasingly focus on fairness and human dignity. Civil society organizations and scholars (many of them faith-based) are shaping these discussions. They do not enforce rules, but they influence public trust. Their concerns resonate with employees and customers alike. Professionals are being asked to explain not only whether AI systems comply with the law but also whether they align with organizational values. Compliance alone no longer resolves these questions.

AI ethics and governance: Where will you draw the line?

Taken together, these forces have altered what is required from leadership. Ethical AI strategy is no longer a discretionary add-on or a future consideration. Leaders must decide how innovation will proceed when regulatory guidance or accepted governance standards are incomplete and consequences are uncertain. They must define what their organizations will tolerate from intelligent systems, and they cannot delegate that responsibility or defer it to regulators. This imperative explains why ethical AI governance now belongs in the boardroom and C-suite. It also explains why leadership judgment matters more than technical expertise.

Excerpted with permission from the publisher, Wiley, from “Boundaries of Tolerance: A Foundational Model for Ethical AI Strategy and Governance,” by Jeffrey N. Saviano. Copyright © 2026 by Saviano Ventures LLC. All rights reserved.

Jeffrey Saviano is a senior lecturer in AI strategy, governance, and ethics at the MIT Sloan School of Management who recently completed a three-year research appointment at Harvard University. He leads the MIT Open Learning course, A New Imperative for AI Governance and Ethics: Winning the AI Race Without Losing Control.

Over more than three decades at EY, Saviano developed and commercialized some 25 businesses and solutions, generating over $400 million in annuity revenue. He led EY’s research collaboration with MIT Connection Science, focusing on enterprise AI strategy, risk, and new governance models for senior executives, and concluded his time at the firm as its emerging technology strategy and governance leader.