Artificial intelligence is transforming executive access to information. But having the answer is not the same as understanding it, defending it, or using it to lead.

Artificial intelligence is rapidly eliminating one of the historical constraints on executive decision-making: access to information. Complex analyses that once required days of research, consultation, and preparation can increasingly be completed in minutes. An executive equipped with sophisticated AI tools or a personalized digital twin can examine operating performance, regulatory developments, industry research, organizational history, and strategic alternatives at a speed and scale previously unavailable to most leaders.

This represents an extraordinary advancement in professional capability. It also introduces an emerging leadership risk that receives considerably less attention than AI accuracy, hallucinations, or automation.

The amount of information available to an executive can now increase much faster than the executive’s ability to understand, internalize, challenge, and credibly command that information.

The result is what I call the Information-to-Command Gap.

An executive can arrive at a meeting with an exceptionally well-prepared, AI-assisted analysis and still be less prepared to lead the discussion than the quality of that analysis might suggest. The information may be accurate, the recommendations technically sound, and the presentation compelling. Yet the executive may lack sufficient personal command of the evidence, assumptions, competing interpretations, or business implications to respond effectively when challenged.

This is a leadership challenge involving the relationship between access to intelligence and the capacity to exercise judgment and influence. I have experienced it myself.

The Executive Meeting That Changed My Thinking

During an acquisition opportunity discussion in my current professional practice, I conducted extensive research to prepare for an important project review with our executive team. My preparation included conventional research methods, supplemented by substantial AI-assisted information identification and retrieval. The combination allowed me to explore a much larger body of information and examine the opportunity more comprehensively than would have been practical using traditional methods alone.

I entered the discussion feeling well prepared.

The executive review was effective. We had a productive conversation about the opportunity, and the research contributed meaningfully to the discussion. However, as the questions became more specific and the conversation moved deeper into certain issues, I began recognizing a gap between the information I had researched and my ability to recall and apply important details in the moment.

I knew I had encountered relevant information during my preparation. I could remember reviewing material that addressed the questions being raised, and I understood its significance at the time. Yet as the discussion became more probing, I could not always readily recall the specific details necessary to provide the depth of explanation I wanted.

The meeting was productive, and the underlying analysis was not necessarily wrong. Nevertheless, I left with the sense that I could have contributed considerably more if I had developed stronger personal command of the extensive research I had conducted.

That realization stayed with me. I had invested significant effort in gathering and analyzing information, but I had not been equally deliberate about preparing to retrieve, connect, and apply the most important elements under questioning.

The problem was not insufficient access to information. I had more information than I could have reasonably assembled through conventional methods alone. The problem was determining which portions I needed to internalize and how to develop the depth of understanding necessary to use them effectively.

That experience led me to begin developing a more intentional approach to maximizing information command. I became less concerned with simply expanding the research available to me and more focused on identifying the essential evidence, examining competing interpretations, and testing whether I could explain the conclusions without relying on the underlying research materials.

The lesson applies beyond AI. Professionals have always faced challenges absorbing large quantities of information. What is changing is the scale and speed at which AI can expand the material available for consideration, often far beyond what an executive has sufficient time to examine personally.

The value of executive preparation is not measured by how much information we have reviewed, but by how effectively we can use the right information when it matters.

Information Is Not Understanding, and Understanding Is Not Influence

For decades, executive advantage has depended partly on privileged access to information. Senior leaders typically had better access to analysts, technical experts, historical records, business intelligence, and professional networks. Their positions gave them visibility into information that others could not easily obtain.

Artificial intelligence is progressively changing that advantage. Information is becoming easier to retrieve, analyze, and synthesize, even as proprietary knowledge, specialized expertise, and organizational context remain important differentiators.

I believe executives increasingly need to distinguish three levels of information capability.

Information Available represents everything the executive can access through personal knowledge, organizational systems, professional networks, conventional research, and AI-enabled resources. This includes historical reports, operating data, regulatory requirements, technical research, industry comparisons, and potential recommendations. AI dramatically expands this layer.

Information Commanded is the portion of that information the executive has personally understood, interrogated, prioritized, and reconciled with experience. It has been transformed from an external information source into an informed professional perspective. The executive understands what the evidence establishes, where uncertainty remains, how competing explanations compare, and which conclusions can be defended.

Information Converted to Influence is where leadership creates organizational value. The executive uses that understanding to communicate clearly, respond to challenges, establish priorities, recommend decisions, build alignment, and move people toward effective action.

These three levels are related, but they are not interchangeable.

Information Available ≠ Information Commanded ≠ Organizational Influence

Figure 1. From Information to Influence: The three levels of executive information capability.

The distinction does not imply that executives must memorize extensive technical details or personally reproduce every calculation. Effective leadership has always relied on specialists, analytical tools, and distributed expertise. Command involves knowing which information matters, understanding the essential reasoning, recognizing the boundaries of one’s expertise, and knowing when additional verification or technical judgment is necessary.

In fact, knowing what does not require personal command is increasingly important. The executive who attempts to absorb everything AI can produce will quickly encounter human cognitive limits. The executive who can distinguish essential understanding from supporting information may gain a substantial leadership advantage.

The Hidden Cost of Cognitive Offloading

The underlying issue has a connection to cognitive science known as cognitive offloading, the use of external resources to reduce the mental effort required to perform a task.

Humans have always used tools to extend cognitive capabilities. Written records reduce the need to remember details. Calculators perform computations. Databases store information, and search engines accelerate retrieval. These technologies allow people to devote attention to activities that may create greater value.

Research by Risko and Gilbert (2016) describes cognitive offloading as a normal feature of human thinking rather than an inherently undesirable behavior.

Generative AI introduces an important difference. Instead of simply storing information or performing predefined calculations, it can participate in activities traditionally considered higher-order cognitive work: synthesis, evaluation, comparison, reasoning, and the generation of arguments or recommendations.

This means executives can delegate not only the collection of evidence but also portions of the intellectual process through which they would ordinarily develop understanding.

The potential to extend human reasoning is among AI’s most valuable characteristics. The concern arises when reliance on AI-produced conclusions reduces the personal engagement necessary to develop independent judgment.

A 2026 study published in Frontiers in Psychology distinguishes between dependent and autonomous cognitive offloading. Dependent offloading involves delegating core thinking to AI, while autonomous offloading uses AI to support thinking while retaining personal cognitive agency. In a study involving 589 university students and early-career knowledge workers, the researchers found that these different patterns of engagement were associated with different perceived cognitive outcomes despite offering comparable immediate benefits (Zhu et al., 2026).

Related research involving 319 knowledge workers found that greater confidence in generative AI was associated with less self-reported critical-thinking effort, while greater confidence in one’s own abilities was associated with more critical engagement. The researchers also identified a shift in cognitive effort toward verification, integration, and oversight (Lee et al., 2025).

These studies do not establish that AI causes deterioration in executive judgment. Much of the research remains exploratory or observational, and the findings are not specific to senior executives. However, they reinforce the importance of examining how AI changes the cognitive activities involved in professional work.

In my experience, preparing a complex executive recommendation has traditionally required considerable time reading reports, examining supporting evidence, comparing interpretations, and constructing an argument. These activities are time-consuming, but they also create familiarity with the subject. The effort involved in developing the analysis contributes to the ability to explain and defend it.

AI can compress much of that process. The efficiency benefit is substantial, but some of the intellectual engagement that previously developed understanding can disappear along with the time required to complete the work.

The fastest route to an answer is not always the fastest route to understanding.

The practical challenge is deciding which cognitive activities should be delegated and which deserve deliberate personal attention. Repetitive research, information organization, and routine analysis are obvious opportunities for efficiency. Evaluating uncertain evidence, reconciling competing conclusions, and determining what an organization should do often require deeper engagement.

Some intellectual friction is unnecessary. Other friction is precisely what develops understanding.

Your Digital Twin Should Challenge You, Not Simply Agree With You

In my work with executive digital twins, I have become increasingly interested in using AI to challenge my own thinking rather than simply accelerate analysis.

A personalized digital twin can become extremely effective at understanding its user’s established perspectives. A system contextualized around an executive’s experience, preferences, professional history, and strategic priorities can generate recommendations that feel immediately familiar and credible.

That familiarity has enormous value, especially when applying accumulated professional experience to new challenges. It can also reinforce existing assumptions if the user does not deliberately request alternative viewpoints.

I now see intellectual challenge as an essential part of effective human-AI collaboration.

When developing a recommendation, I can ask the system to identify its weakest assumptions, construct a credible argument against my preferred conclusion, or explain what evidence would justify changing direction. Before an important executive discussion, I can use the digital twin to adopt the perspective of a skeptical CEO, board member, operations leader, or regulator.

The resulting conversation can be uncomfortable, particularly when the AI identifies a weakness in an argument I believed was well developed. But that is precisely where much of the value lies.

For example, an executive preparing to recommend a significant capital investment could ask AI to construct the strongest argument for delaying the expenditure, identify evidence that might favor another alternative, and explain which assumptions would be most difficult to defend under questioning.

This is a useful way to move from having a persuasive recommendation to having a recommendation that has survived meaningful intellectual challenge.

Research on cognitive ownership provides additional context. Su et al. (2026) examined associations between habitual AI dependency and concerns about whether AI-assisted work still felt intellectually attributable to the person producing it. The study measured perceived concerns among pre-service teachers rather than demonstrated deterioration in professional capability, but it raises an interesting question about maintaining ownership of one’s thinking.

For an executive, intellectual ownership is demonstrated through the ability to assess evidence, challenge assumptions, and defend conclusions.

A valuable digital twin should be capable of thinking with you, but it should also be capable of thinking against you.

The Executive Command Protocol: Turning Information into Leadership

My acquisition-review experience prompted me to think more deliberately about how to prepare for consequential executive discussions. I began developing an approach that takes advantage of AI’s analytical capacity while ensuring that the most important information becomes part of my own working understanding.

I call this the Executive Command Protocol.

The approach is built around four repeating phases that connect information gathering, personal understanding, leadership interaction, and subsequent learning.

Figure 2. The Executive Command Protocol: AI Expands → Human Compresses → Human Commands → AI Expands Again.

Phase 1: AI Expands

The process begins by taking advantage of AI’s ability to expand the information environment.

The executive uses AI alongside conventional research to retrieve and organize relevant information, examine operating data, compare historical experience, identify potential patterns, explore alternative explanations, and anticipate stakeholder questions.

During this phase, I want the analysis to be broad enough to reveal issues that might otherwise be overlooked. This includes actively searching for contrary evidence rather than limiting research to material supporting an initial position.

Source verification remains important. AI-generated interpretations should be distinguished from documented facts, and the reliability of critical evidence must be established before it informs consequential decisions.

Phase 2: Human Compresses

This is the most important transition in the protocol and the area where my own experience revealed a need for greater discipline.

Once the research has expanded, the executive needs to determine which information deserves personal command. The goal is to build sufficient understanding to exercise judgment and participate effectively in the discussion, without attempting to memorize every fact uncovered during the research.

I organize this process around four questions:

What matters? Which facts, risks, assumptions, or uncertainties fundamentally affect the decision?

What do I believe? After considering the evidence and alternative interpretations, what is my professional conclusion?

What do I recommend? What course of action should be taken, and why is it preferable to the alternatives?

What am I prepared to defend? Can I explain the reasoning, evidence, tradeoffs, and uncertainties when challenged?

To make these questions practical, I separate the research into three categories.

Must Command includes the central conclusion, critical supporting evidence, material uncertainty, business implications, recommended decision, and strongest counterargument. These are issues I should be able to explain and defend without relying heavily on reference material.

Should Recognize includes supporting trends, secondary operating information, historical context, stakeholder perspectives, and related initiatives. I need to understand their relevance and recognize when they become important, but precise recall is less essential.

Available on Demand includes detailed statistics, extensive historical records, secondary calculations, technical references, and background information that can be reliably retrieved when needed.

This classification is particularly useful because an executive discussion rarely follows a perfectly predictable sequence. Details that initially seem secondary can become central when a participant raises a new issue. Good preparation therefore includes knowing where supporting evidence resides and how to retrieve it efficiently.

I have also begun thinking about information command as something that should be tested rather than assumed.

One practical approach is to prepare a short, decision-focused summary identifying the recommendation, essential evidence, assumptions, uncertainties, and strongest alternative. Each consequential claim should be traceable to its supporting source. The summary becomes a guide for developing understanding rather than another lengthy briefing to memorize.

The second step is to use AI to challenge that understanding. Have the digital twin ask progressively more detailed questions, beginning with the business implications and moving into the supporting evidence. Ask it to introduce competing interpretations and challenge conclusions that rely on uncertain assumptions.

Then rehearse explaining the recommendation without the briefing in front of you. When an answer requires specific evidence that cannot be recalled, return to the relevant material, examine it, and determine whether it belongs in the Must Command category or can reasonably remain available for retrieval.

This method combines the efficiency of AI-assisted research with deliberate cognitive engagement. It helps distinguish information that feels familiar because I have read it from information I understand well enough to use.

Phase 3: Human Commands

During the leadership interaction, preparation becomes visible through the executive’s ability to listen, evaluate new information, explain tradeoffs, respond to challenges, and adapt recommendations as the discussion evolves.

The AI may remain available as a reference or analytical resource. There is nothing inherently deficient about consulting supporting information or acknowledging that a particular detail requires verification.

Strong information command allows the executive to recognize when additional information is necessary without losing sight of the central issue.

This matters because executive discussions often move quickly between strategic considerations, operational consequences, financial implications, and technical details. A leader needs sufficient command to maintain the integrity of the recommendation while responding constructively to different perspectives.

Phase 4: AI Expands Again

After the discussion, AI can help capture decisions, identify follow-up actions, organize new information, and preserve lessons for future work.

The questions raised during the meeting also become useful evidence about the quality of preparation. Were there recurring gaps in understanding? Did a competing explanation emerge that had not been adequately examined? Did the discussion reveal information that should have been classified as essential?

These observations can improve the next cycle of research and preparation.

Where confidential business information is involved, including potential acquisitions, AI use must remain within authorized systems and established information-security requirements.

AI Expands → Human Compresses → Human Commands → AI Expands Again

The protocol is intended to improve professional capability and leadership effectiveness. Its ultimate value should be measured through better-informed decisions, stronger executive discussions, and improved organizational outcomes.

Developing Information Command Across the Executive Team

Although I began developing this approach in response to my own experience, its potential extends beyond individual preparation. Executive teams regularly make consequential decisions based on information assembled by different functions, advisors, and analytical systems. Each participant may have substantial expertise in their own discipline while possessing only a partial understanding of the broader decision.

The Executive Command Protocol can provide a common preparation discipline for leadership teams. Before a major investment, acquisition, or operational review, participants can agree on the essential evidence, material assumptions, principal uncertainties, and decisions requiring resolution. AI can help identify gaps and generate competing interpretations, while structured discussions allow leaders to challenge one another’s reasoning and establish a shared understanding of the issues.

This also creates an opportunity for leadership development. Rather than evaluating managers primarily on the quality of their presentations, senior executives can develop their teams by probing how recommendations were formed, asking what evidence could change their conclusions, and examining how decisions performed against expectations. Reviewing the questions that exposed gaps in understanding can improve future preparation and reinforce independent judgment.

The objective is not for every executive to command every technical detail. It is to ensure that the leadership team collectively possesses the understanding necessary to make sound decisions, knows where specialized expertise is required, and can explain the reasoning behind its recommendations.

Used consistently, this discipline could make AI-supported preparation part of developing stronger executive judgment across the organization.

From Human in the Loop to Human in Command

The broader discussion of responsible artificial intelligence has long emphasized human oversight. The European Commission’s High-Level Expert Group on Artificial Intelligence identified human-in-the-loop, human-on-the-loop, and human-in-command approaches in its 2019 Ethics Guidelines for Trustworthy AI.

In that established governance context, Human in Command refers to meaningful human oversight of AI systems, including authority over their use, discretion, and ability to override decisions.

I believe the concept has an important additional application to executive leadership: maintaining personal command of the information and reasoning that AI systems help produce.

A leader can participate in an AI-assisted decision process without fully understanding its conclusions. Reviewing or approving a recommendation does not necessarily demonstrate independent judgment. Effective executive command requires sufficient understanding to question the analysis, recognize uncertainty, and take responsibility for the recommendation.

The distinction becomes particularly important as digital twins move beyond information retrieval into increasingly sophisticated analysis, decision support, and potentially autonomous actions.

For executives, my preferred formulation is straightforward:

AI proposes. I evaluate. I decide. I own the outcome.

This does not mean the executive must personally possess every technical skill represented in the analysis. It means that delegating analytical work to AI does not transfer the leadership responsibility to exercise appropriate judgment.

Why This Matters for EHS and Operational Integrity

The consequences of inadequate information command can be particularly significant in environmental, health, safety, and operational integrity functions. Decisions frequently involve serious hazards, regulatory obligations, capital investments, process controls, emergency preparedness, and potentially irreversible consequences.

Consider an EHS professional presenting an AI-generated risk assessment recommending a substantial investment in engineered safeguards.

The analysis may identify a high-consequence scenario and develop a compelling business case. Yet an operations executive might reasonably ask whether the probability assumptions are credible, whether existing safeguards were properly evaluated, or whether another intervention could provide comparable risk reduction with less operational disruption.

The professional needs to understand the underlying hazard, potential consequences, safeguards, exposure scenarios, and uncertainties. The recommendation must also account for the real operating environment in which the controls will be implemented.

This is closely related to a familiar EHS distinction: the availability of a control does not establish its effectiveness.

A written procedure can exist without being followed. Training can be completed without producing competence. A safeguard can be installed without performing reliably under actual operating conditions.

Similarly, an organization can possess an excellent AI-generated risk assessment without having developed the professional understanding necessary to implement and sustain effective controls.

EHS professionals recognize the importance of verifying whether critical controls actually work. We should apply similar discipline to AI-supported professional judgment by testing whether those presenting recommendations can explain the assumptions, identify material uncertainty, and connect the analysis to the operational conditions being evaluated.

I expect AI to significantly improve our ability to identify patterns, examine complex operating conditions, and recognize emerging risks. Realizing those benefits will require professionals who understand both the capabilities and the limitations of the technology.

The greater the analytical sophistication, the more important that understanding becomes.

The New Executive Differentiator

The implications extend well beyond EHS.

Across finance, operations, engineering, sustainability, supply chain, legal affairs, and enterprise risk management, AI is changing the economics of knowledge work. Information that was historically difficult or expensive to obtain can increasingly be made available with relatively little effort.

An executive who can produce a comprehensive strategic analysis in minutes may no longer possess an unusual advantage if other executives have access to similar capabilities. Greater value may come from interpreting that analysis, recognizing where the evidence is weak, understanding the organizational context, and exercising judgment when the appropriate course of action is uncertain.

Consider the commercial implications in a potential acquisition. Research may uncover extensive information about operating performance, regulatory obligations, environmental liabilities, capital requirements, and integration challenges. However, creating business value depends on distinguishing which findings are material to the investment thesis and understanding how they could affect valuation assumptions, transaction terms, future capital commitments, or post-acquisition integration priorities.

An executive who commands those relationships can help the leadership team recognize risks that might otherwise be underestimated, challenge assumptions behind projected returns, and identify opportunities that may not be apparent from the information alone. The same principle applies to major capital projects, market-entry decisions, restructuring, and other strategic investments. The ability to convert information into sound judgment has direct implications for enterprise value creation.

This introduces an interesting organizational performance concern.

Companies can become dramatically more efficient at generating reports, analyses, presentations, and recommendations without necessarily improving the capabilities of the people responsible for interpreting and acting upon them.

If organizations measure AI adoption primarily through time savings and output volume, they may overlook whether professional understanding, independent reasoning, and decision quality are also improving.

There is an opportunity to approach AI implementation differently. Organizations can deliberately design AI-supported workflows to strengthen professional competence, incorporate intellectual challenge, improve knowledge retention, and develop better decision-making skills.

The executive digital twin is particularly promising in this respect. It can preserve accumulated experience, introduce alternative perspectives, connect current problems to historical decisions, and provide realistic opportunities to test professional reasoning.

My own experience has made me more conscious of the need to develop both sides of this capability. I want my digital twin to become increasingly effective at helping me research, analyze, and challenge complex issues. At the same time, I want to become more effective at identifying what I personally need to command and developing the understanding necessary to use that information.

The opportunity is to make technology and human capability advance together.

The Executive Command Test

Before an important executive discussion, I have come to value a simple test: temporarily put the AI-generated materials aside and determine whether I can explain the issue without them.

Can I identify the evidence supporting my conclusion and distinguish what is established from what remains uncertain? Do I understand the strongest competing explanation? Can I describe what would cause me to change my position? Can I explain the recommended action, its business implications, and the principal risks?

The test becomes more valuable when another person—or an AI acting as a skeptical stakeholder—asks follow-up questions that were not anticipated during preparation.

It is not a memory examination. Executives should continue relying on technical specialists, reliable reference materials, and appropriate analytical tools. The purpose is to determine whether the essential reasoning has become part of their professional understanding.

When the answers reveal gaps, additional preparation can be targeted rather than simply expanding the research again.

This is especially important because sophisticated AI-generated material can create a misleading sense of preparation. A clear, comprehensive briefing may make an issue feel familiar even when the underlying evidence and reasoning have not been sufficiently internalized.

Recognizing that difference before entering the meeting is far preferable to discovering it under questioning.

Putting the Executive Command Protocol into Practice

To help professionals apply this approach, I developed the Executive Information Command Drill, a reusable AI prompt designed to turn completed research into demonstrated executive readiness. Rather than generating another summary or presentation, the tool challenges users to identify what they must personally command, defend their conclusions against increasingly difficult questions, and recognize gaps in their understanding before entering an important discussion.

The drill follows the Executive Command Protocol and can be applied to strategic planning, acquisition reviews, capital investment proposals, operational risk assessments, and other consequential decisions.

[Access the Executive Information Command Drill — AI Prompt]

In an Information-Rich World, Command Becomes the Scarce Resource

Artificial intelligence and executive digital twins offer professionals an unprecedented opportunity to extend their capabilities. They can bring decades of accumulated knowledge into current decisions, identify relationships across large bodies of information, challenge assumptions, and help leaders prepare for complex issues with greater breadth and depth.

I believe these technologies will become indispensable to effective executive leadership. Their value should be measured through the quality of the decisions and actions they enable, not simply the volume of work they produce.

My acquisition-review experience revealed a gap in my own approach to preparation. I had conducted extensive research and participated in a productive executive discussion, yet I recognized afterward that I could have contributed more. There was important information I had encountered but could not readily bring into the conversation when more detailed questions arose.

The experience changed how I approach AI-assisted preparation. I now place greater emphasis on deciding what I need to command, examining the evidence behind important conclusions, and testing my understanding before consequential discussions.

The lesson is becoming more relevant as AI continues to advance. Executives will have access to increasingly sophisticated analysis and recommendations, but that access will not eliminate the need for personal judgment, intellectual discipline, or leadership capability.

I remain highly optimistic about what artificial intelligence will enable. The opportunity to extend professional knowledge and improve executive decision-making is extraordinary. Realizing that opportunity requires us to become as deliberate about developing human understanding as we are about expanding technological capability.

Let AI expand what you can know. Be deliberate about what you must understand. Use the technology to challenge and strengthen your judgment, and develop the command necessary to apply that knowledge when it matters most.

In an information-rich world, command becomes the scarce resource.

References

High-Level Expert Group on Artificial Intelligence. (2019). Ethics guidelines for trustworthy AI. European Commission. https://digital-strategy.ec.europa.eu/en/library/ethics-guidelines-trustworthy-ai

Lee, H.-P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, Article 1121, 1–22. https://doi.org/10.1145/3706598.3713778

Risko, E. F., & Gilbert, S. J. (2016). Cognitive offloading. Trends in Cognitive Sciences, 20(9), 676–688. https://doi.org/10.1016/j.tics.2016.07.002

Su, J., Wang, Z., Qiang, Z., & Wang, Z. (2026). Is it mine or not mine? Cognitive offloading to generative AI and concern over the erosion of cognitive ownership. Journal of Intelligence, 14(8), Article 182. https://doi.org/10.3390/jintelligence14080182

Zhu, Q., Li, X., Dong, Y., Chang, P., & Fan, M. (2026). Not all cognitive offloading is equal: Distinguishing dependent and autonomous offloading to generative AI. Frontiers in Psychology, 17, Article 1878629. https://doi.org/10.3389/fpsyg.2026.1878629

Concept Elements Poster