The Central Question
What happens when corporate boards and C-suite leaders are asked to integrate one of the most transformative technologies in history, artificial intelligence, into their core strategies without the frameworks or regulatory clarity needed to lead responsibly?
The Core Thesis
AI is advancing faster than regulation, and today’s leaders bear the responsibility for ensuring its safe and trustworthy use. Ethical AI governance is a defining test of leadership and organizational purpose.
The book introduces the Boundaries of Tolerance Model, a practical governance framework developed at Harvard University’s Edmond & Lily Safra Center for Ethics, and advanced in my teaching at MIT. The framework helps organizations and individuals:
• Define acceptable risk thresholds for AI systems
• Translate abstract ethical principles into concrete oversight practices
• Clarify the ethical boundaries of what organizations, and the humans involved, will and will not tolerate in their AI systems
Why It Matters
The risks of AI are already visible: biased algorithms, reputational damage, shareholder litigation, and loss of public trust. Fiduciary standards from landmark Delaware cases (Caremark, Boeing, McDonald’s, and others) demonstrate that corporate boards and business executives are legally accountable for mission-critical risks, including those grounded in AI.
Most current books on AI focus on technology-enabled strategy, governance, risks, or ethics in isolation. Boundaries of Tolerance integrates these perspectives into a unified governance model grounded in fiduciary duty, applied ethics, leadership, and organizational design.
What Makes the Book Compelling
Boundaries of Tolerance reframes AI governance as a new frontier of corporate leadership. Through case studies and original research, it offers leaders the tools to embed ethics into strategy and align innovation with societal values.
The book presents a structured, actionable roadmap to navigate competing tensions:
• Innovation and oversight
• Speed and accountability
• Transparency and privacy
• Enterprise strategy and ethical integrity
The Model in Action
At the heart of the book is the Boundaries of Tolerance Model, a practical framework for ethical AI governance. It equips boards and senior executives with structured tools to define their ethical boundaries and integrate them into enterprise oversight. The Model is built on three pillars:
-
Ethical Maturity Stack: Charts an organization’s evolution from non-compliance at the base to ethical vanguard at the summit.
-
Performance Measures (Enterprise and Systems): Internal assessment tools that help boards and executive teams determine how ethical principles are operationalized across the business and its AI systems, and determine where the company sits on the Ethical Maturity Stack.
-
Leading and Emerging Practices: Offers best practices and next generation tools business leaders can adopt today, such as establishing a board-level technology or AI committee, defining “guard bands” of safety, or adopting principles of ‘Prudent Vigilance’ as a balanced innovation / risk approach, calling for business enterprises to continuously monitor and adapt their AI systems in response to evolving risks and uncertainties.
Together, these pillars move organizations toward measurable progress in ethical AI maturity.
Business AI Ethics Research Initiative
The Boundaries of Tolerance model was developed by Jeffrey Saviano with research support from the Business AI Ethics Research Initiative, which he launched in August 2023 during his appointment at the Edmond & Lily Safra Center for Ethics at Harvard University.
The initiative focuses on advancing practical understanding of how organizations can responsibly govern artificial intelligence as it becomes embedded in business operations. The research examines the intersection of AI deployment, regulatory developments, and corporate governance responsibilities.
A central focus of the initiative is exploring how boards and executives can effectively oversee AI-related risks while still enabling innovation. The research team develops practical insights and frameworks that help organizations translate high-level ethical principles into actionable governance practices.
The initiative’s research – including analysis of global legal and policy developments, technology trends, corporate case studies, and documented AI incidents – supports the ongoing development and application of the Boundaries of Tolerance model.
The initiative brings together a growing, cross-disciplinary research team with expertise spanning law, business and finance, technology, and public policy across multiple jurisdictions. The team contributes research, performs analysis, and shares insights through publications – including the blog AI Ethics for Business Leaders – as well as presentations and discussions with both professional and academic audiences.
The Research Team
Vija Kalniņa, PhD
Research Director, Scientific Editor
Vija Kalniņa is a legal expert and former judge at the Court of Economic Affairs in Latvia who leads the BAIE research team. She contributes research and analysis to the Boundaries of Tolerance project, helps shape its evolving research agenda, and presents the team’s findings to diverse professional audiences.
She holds a PhD in European Union (EU) technology law and has more than 15 years of experience as a lawyer and judge working on complex business law matters, including white-collar crime and commercial disputes.
Vija is an advocate for AI literacy and responsible AI governance. She teaches and speaks about AI ethics, law, and governance to a wide range of audiences – including law students, judges, EMBA students, and other professionals – and writes about AI technology and law, including EU AI policies, and their impact on legal professionals and business organizations.
She joined the BAIE team in December 2024 during her Hubert H. Humphrey (Fulbright) Fellowship (2024-2025) at Boston University’s Questrom School of Business. She continued her work with the team beyond the program, driven by a strong commitment to advancing innovative approaches to ensuring trustworthy AI and helping businesses strengthen responsible AI governance in a rapidly evolving legal and policy environment.
Lily Noyes
Research Assistant
Lily is a senior at the University of Pennsylvania pursuing a double major in Physics and Classical Studies. She joined the BAIE team in May 2025, where she contributes research into the evolving global AI landscape. Her work includes international policy analysis, the examination of AI incidents and real-world use cases, and research into the broader societal impacts of AI systems. She communicates these findings through delivering presentations and publications for both corporate and academic audiences.
Lily’s interests center on the intersection of science, technology, and public policy with a particular focus on how AI can be responsibly developed and applied to address complex social challenges. In the coming year, Lily will be pursuing a Master’s degree in Social Data Science at the University of Oxford. Through this program, she plans to deepen her understanding of how data and AI systems shape decision-making, while developing the technical skills to study and guide the responsible use of these technologies.
Ryan Wettre
Research Assistant
Ryan Wettre is a senior at Northeastern University, pursuing a bachelor's degree in business administration with a concentration in Finance and a minor in Mathematics. Ryan joined the BAIE team in June of 2025, and his research work focuses on the intersection of AI governance and evolving regulatory landscapes for AI systems. He has worked with the team to disseminate research findings to wider audiences through publications focused on US policy analysis.
Ryan’s background has been shaped by experiences in both business and law, giving him a perspective that highlights the growing overlap between technological innovation, corporate decision-making, and legal governance. This interdisciplinary perspective has built his interest in how legal and institutional frameworks can guide the responsible development and deployment of AI.
Ryan will be attending law school in the coming year, where he hopes to further explore the legal and governance challenges posed by artificial intelligence, with a particular interest in regulatory frameworks, corporate oversight, and the accountability of organizations deploying AI systems.
Colton Mikolajczyk
Research Assistant
Colton is currently a candidate for a Master’s degree in Business Analytics at the Massachusetts Institute of Technology (MIT). Prior to MIT, he attended Rensselaer Polytechnic Institute (RPI), where he graduated in 2025 with degrees in Mathematics and Business Analytics. He works as a research assistant with the BAIE team, where he helps translate the team’s research into clear and accessible visual communication through artwork and visual designs. He also supports the team’s broader research efforts by contributing insights that advance its objectives, including helping ensure alignment with relevant regulations and standards, compiling technical documentation, and examining how AI applies in practice across sectors such as finance.
His interests in artificial intelligence center on regulation and the development of meaningful safeguards for AI use within companies. He is interested in ensuring that people remain aware of how they use AI and of the effects that AI can have on coworkers, organizations, and the consumers of their products. Colton joined the team in February 2026 and is expected to graduate from MIT in August 2026. Following graduation, he hopes to work in industry, where he can apply principles aligned with the book’s perspective on responsible AI use.