Artificial intelligence is becoming increasingly embedded in business operations, strategic planning, customer service, analysis, and decision-making. As organizations adopt increasingly capable AI systems, an important leadership question is emerging: where should automation end and human judgment begin? Jason Norbeck examines why human expertise and accountability remain important even as businesses expand their use of artificial intelligence.
AI can process large quantities of information, identify patterns, generate content, and support decisions at a speed that traditional processes cannot match. However, speed and computational capability do not eliminate the need to understand context, evaluate consequences, challenge assumptions, and take responsibility for important decisions. Jason Norbeck emphasizes that responsible AI adoption requires organizations to determine where human review provides essential value.
Morgan Stanley’s August 2026 analysis of corporate AI governance found that human oversight remains central to how companies are managing AI risks. Among 200 executives involved in AI governance at global corporations, 41% identified separate human review in higher-risk situations as the most important guidance for employees using AI.
AI Can Support Decisions Without Replacing Responsibility
Jason Norbeck explains that one of the most important distinctions in AI adoption is the difference between assisting a decision and owning a decision. An AI system may provide an analysis, recommendation, prediction, or generated response, but an organization still needs to determine who is accountable for acting on that information.
This distinction becomes especially important when AI is used in areas involving financial decisions, hiring, customer relationships, compliance, cybersecurity, or other consequential business activities. Leaders should establish clear decision rights so employees understand when AI recommendations can be accepted and when additional human evaluation is required.
Jason Norbeck notes that automation should therefore be designed around organizational responsibility rather than treated as a substitute for it. An organization cannot simply attribute an undesirable outcome to an algorithm and consider the matter resolved. Leadership remains responsible for establishing the conditions under which AI is deployed and monitored.
Human Expertise Provides Context AI May Miss
AI systems can analyze information efficiently, but Jason Norbeck emphasizes that organizational decisions often depend on context that may not be fully represented in the data available to an AI system.
Experienced professionals can recognize unusual circumstances, question incomplete information, consider competing priorities, and understand consequences that may not be obvious from a dataset. Human expertise can therefore provide an important layer of interpretation when AI outputs are used to inform significant decisions.
Morgan Stanley’s 2026 research found that 68% of S&P 500 companies had at least one non-executive director with identified AI skills, while one-quarter had three or more. The research indicates that boards are developing deeper AI expertise as they seek to oversee the opportunities and risks associated with adoption.
Jason Norbeck explains that this does not mean every leader needs to become an AI engineer. It does mean executives and directors need enough understanding to ask meaningful questions about how AI is being used, what limitations exist, and where human expertise should remain involved.
Higher-Risk Decisions Require Greater Oversight
Not every AI application carries the same level of risk. Jason Norbeck notes that organizations should distinguish between routine uses of AI and applications that could produce significant financial, legal, operational, employment, privacy, or reputational consequences.
A system that helps summarize an internal document may require a different level of oversight than an AI application influencing a high-impact business decision. Establishing risk-based review can allow organizations to benefit from automation while concentrating human resources where judgment is most important.
Morgan Stanley found that 69% of surveyed executives said their companies provide guidance requiring AI outputs to undergo separate human review in higher-risk situations. Human review was the most frequently selected response when executives were asked which responsible-AI guidance was most important.
Jason Norbeck emphasizes that effective human oversight should be substantive rather than symbolic. Employees conducting reviews need enough authority, expertise, time, and access to relevant information to question an AI-generated recommendation when necessary.
Leadership Must Understand AI Limitations
Jason Norbeck explains that responsible AI leadership requires understanding not only what an AI system can do but also where it may fail. AI outputs can contain errors, reflect limitations in training data, or produce confident responses that require additional verification.
Leaders should therefore establish expectations around validation and escalation. Employees need to know which outputs require independent verification and what steps to take when an AI-generated result appears incomplete, inconsistent, or inappropriate.
Morgan Stanley’s research found that data risks—including privacy, cybersecurity, and discrimination—were identified as material sources of AI risk by 90% of surveyed executives, with 56% identifying data risks as the top current concern. Jason Norbeck notes that these concerns reinforce the importance of maintaining human review when AI systems operate in sensitive environments.
AI Should Strengthen Expertise Rather Than Eliminate It
The most productive approach to AI does not necessarily involve replacing human expertise. Jason Norbeck explains that organizations can instead use AI to help professionals work more efficiently while preserving human responsibility for interpretation and important decisions.
AI can summarize large volumes of information, identify patterns, automate repetitive processes, and generate preliminary analysis. Employees can then devote more time to complex judgment, relationship management, problem-solving, strategic thinking, and other activities that require human expertise.
This approach also changes the leadership challenge. Jason Norbeck notes that executives need to think about how jobs and responsibilities should be redesigned when AI becomes part of everyday workflows. Simply adding an AI tool to an existing process may not produce the intended results if employees do not understand how the technology changes their responsibilities.
Accountability Must Remain Clear
Jason Norbeck emphasizes that human judgment has limited value if accountability is unclear. Organizations need to establish who owns AI systems, who monitors performance, who reviews significant outputs, and who has authority to intervene when problems arise.
Morgan Stanley reported that around 90% of surveyed executives said their organizations had defined or were in the process of defining where responsibility for AI governance and risk management sits. However, fewer than half said responsibility was fully allocated across corporate functions, senior executives, and the board, while another 45% said the work remained in progress.
These findings illustrate why AI governance remains an evolving leadership responsibility. Jason Norbeck explains that assigning ownership can help prevent situations in which multiple departments assume another team is responsible for monitoring an AI system or responding to an unexpected outcome.
Human Judgment Builds Trust in AI Adoption
Trust is another reason human judgment remains important. Jason Norbeck explains that employees, customers, and other stakeholders may be more comfortable with AI when they understand that meaningful human oversight exists.
Transparency about how AI is being used can help organizations communicate the role of automation without suggesting that technology has replaced organizational responsibility. Leaders can explain where AI contributes to a process, where human review remains required, and how concerns can be raised.
Jason Norbeck notes that this approach can also strengthen internal confidence. Employees who understand their role in reviewing and challenging AI outputs may be better positioned to identify problems before they affect customers, operations, or organizational performance.
The Future of AI Leadership Remains Human-Centered
Artificial intelligence will continue changing how organizations work, but Jason Norbeck concludes that greater automation does not eliminate the need for human leadership. Instead, it changes where human judgment is most valuable.
Leaders need to determine which decisions can be automated, which require human review, and which should remain primarily human-led. They also need to establish governance structures that make responsibility clear and ensure that employees have the skills required to work effectively with AI.
Morgan Stanley’s 2026 research shows that companies are continuing to develop AI expertise, define governance responsibilities, and establish controls as adoption expands. Jason Norbeck emphasizes that these efforts demonstrate an important principle: successful AI adoption depends not only on technological capability but also on the quality of human oversight surrounding it.
The role of human judgment is therefore unlikely to disappear as AI becomes more capable. Jason Norbeck explains that judgment may instead become more focused on context, ethics, accountability, risk, and consequences—the areas where organizational leadership must remain responsible for the outcomes of its decisions.
Review the Morgan Stanley analysis for additional information about AI governance, board expertise, human oversight, data risks, and corporate AI adoption.
Subscribe to the JasonChristopherNorbeck.com newsletter for additional insights on ethical leadership, responsible AI, corporate governance, organizational culture, and digital transformation.