Who should be in charge of AI among the C-level executives in the company?

2026-05-09
图片


2026year1Month, a family of wealth500The Strong Insurance CompanyCEOConvene its executive team to discuss the companyAIOwnership of strategic leadership. The Chief Information Officer believes the answer is obvious: autonomousAIThe system should be under her management. The Chief Operating Officer countered that autonomous labor belongs to the operational category. The Chief Financial Officer pointed out that there areAIThe system is making underwriting decisions that directly affect profit and loss. The Chief Risk Officer observed that the autonomous decision-making system is a significant risk exposure. The Chief Human Resources Officer willAIThe intelligent body is functionally equivalent to the existence of employees, which falls within its scope of responsibility. The Chief Data Officer reminds everyone that the success of the whole thing depends on the correct permissions and data access, which belongs to her field.


The meeting ended in failure. A few months later, the company drafted a proposal similar to that of a CEOAIThe recruitment requirements for official positions have included Dell, Pfizer, PwC, UBS, andExpediaWaiting for the company's ranks.


Similar scenes are happening in various placesCPerformed within the senior management team. autonomousAI——An autonomous system with goal orientation, situational awareness, the ability to invoke other software tools, make complex decisions, and take action——A territorial battle is being sparked at the top.Who controlsAI”This question isAIThe key organizational structure issue at the dawn of the era will affect the company's strategy, investment level, and the distribution of power and influence among leaders.


This competition for influence is not a new issue. Sociologist Andrew·Abbott in its1988One of the most insightful frameworks for understanding this issue was proposed in the landmark work "Career Systems" of [year]. Abbott pointed out that occupational groups are trapped in permanent competition for control over which areas of work, and significant technological or social disruption is the moment when these boundaries are redefined. Apply his theory toCA high-level executive team can understand why the current situation isAIThe fierce competition has also pointed out a practical path for leaders to break free from territorial disputes and establish a truly effective organizational structure.



Jurisdiction competition


Abbott believed that various professions exist in an interdependent system, constantly competing with each other for control of the workplace. A profession claims to have certain types of jobs that fall under the jurisdiction of its members. It can be seen as their professional territory. The problem is that the control of these territories is not permanently fixed, but rather unclear responsibilities between different professionsFuzzy zoneIt is determined through repeated competition and negotiation and may change ownership at any time. Practicing nurses and physician assistants now undertake primary healthcare tasks that were once only the responsibility of general practitioners. Certified public accountants and tax lawyers compete in complex tax planning, just as lawyers and financial advisors are involved in estate planning business. Social workers compete with psychologists, psychologists compete with psychiatrists, and the latter once competed with neurologists in treating mental health issues.


Competition for jurisdiction has long existed within organizations. Main corporate functions——Supply chain, operations, research and development, information systems, product management, human resources, finance, marketing, sales——Forming an interdependent work system, each function claims to have certain decision-making power. In mature organizations during stable periods, the division of labor between various functions is stable.


However, in times of great change, stability will give way to turbulence. We have now entered such a period. This is like the decline of an empire: stakeholders perceive that the corporate landscape will be redrawn, with each function providing reasons why this territory should belong to them. The boundaries between functional leaders that have been solidified for many years suddenly become negotiable.autonomousAIWe are creating a completely new and unclear field of work, while also changing decisions that only a year ago clearly belonged to specific functional areas.


The competition for jurisdiction has resurged, and this contest has just begun.



Why autonomous AI is different from the past


In a sense, this is not new; Previous technological changes have also sparked local debates about territory.


For example, enterprise resource planning systems(ERP)The emergence of this has triggered the involvement of business process managers andITThe tense relationship between departments. In the end“Control level "——The power to establish rules, permissions, and layout logic——Usually classified as Chief Information Officer, but the relevant business area leaders have important speaking power. Cloud computing initially seemed to aim to decentralize computing power, thereby eliminating the need for centralized control. By transforming the high capital expenditure of purchasing hardware into low operating expenses of (initial) leasing, business department heads only need to swipe their credit cards to activate some software. However, this freedom has been proven to be temporary. With the surge in cloud spending, security risks and compatibility issues have emerged, forcing CIOs to step back and eliminate the shadows quietly breeding in business departmentsITOrganize and reintegrate into the track of centralized control.


autonomousAIIt will bring about more thorough changes. Returning to the example of an insurance company that handles claims independentlyAIProxy is actually a technical system (composed ofCIOResponsible), also an operational workflow (byCOOResponsible), or the focus of risk management (byCROResponsible for making financial decisions that have an impact on profits and losses (by)CFOResponsible for), a new type of labor force (composed ofCHROResponsible), as well as data users (byCDOResponsible for). The decisions and actions of this intelligent agent span across multiple functional jurisdictions and are at the core of most or even all of these domains. andSalesforceWorkdayandNetSuitethis type ofSaaSDifferent platforms, autonomousAIBlurred“Tools "And“Work "The boundary between them. As Silicon Valley observers often say today: in autonomous modeAIIn the era, software itself is a service.


In short, intelligent agents themselves are underwriting teams, or at least the initial drafts they are responsible for judging. Therefore, whoever is responsible for managing these intelligent agents——Including their roles, permissions, constraints, evaluations, upgrades, and budgets——They will increasingly be able to control the content and methods of company decision-making.


图片



Competitors and their claims


Abbott discovered that,The competition for territory between professions is essentially a struggle for linguistic power. Who can successfully frame a new problem within their own professional knowledge system and make the outside world believeWho but me?Whoever wins jurisdiction.This is in today'sCIt is happening in the senior management team.


The Chief Information Officer's proposition is based on infrastructure.AIIntelligent agents must be built, integrated into existing systems, resistant to new attack surfaces, and capable of large-scale maintenance. Creating and running these systems requires professional technical skills and management.


The Chief Operating Officer will be autonomousAIRegarded as a part of the company's core operations. Intelligent agents will exist in workflows, and workflows are operations. The Chief Operating Officer's proposition is that the purpose of intelligent agents is to execute business processes, and optimizing processes is the reason for establishing the position of Chief Operating Officer.


The Chief Human Resources Officer's proposition is a positive challenge to the traditional division of powers. Intelligent agents are a new category of employees in terms of functionality. This meansAIIntelligent agents need to go through a complete organizational process: they not only need to be input into company knowledge, immersed in corporate culture, but also supervised and evaluated during deployment; More importantly, the collaborative mode with human colleagues must be carefully designed to ensure that it can positively impact team morale and consolidate culture, rather than eroding the organization's long-term accumulated knowledge assets. Moreover, they will complicate various aspects of human resource management and bring many challenges to the human resources department.


The Chief Financial Officer's position is based on economic accountability. Intelligent agents will ultimately change approval, pricing, asset utilization, and company spending, thereby affecting financial figures. The intelligent agent authorized to make compensation, sign contracts, make credit decisions or refunds on behalf of the company shall bear financial responsibility.


Risk and legal leaders view autonomous systems as a new source of failure risk, liability risk, and regulatory risk. As we have already seen, whenAIWhen an intelligent agent refuses a loan or job application, rejects insurance claims, or becomes involved in a traffic accident, it may trigger litigation and regulatory review.


Finally, the Chief Data Officer pointed out that all other functional claims rely on data. Due to the fact that intelligent agents reason based on information, the reliability of their output depends on the quality of the input information. Who, under what governance rules, and what data can be accessed will determine whether the intelligent agent can work properly, let alone whether its work is secure.



From 'Who owns AI?' to 'Who has what decision-making power?'


These claims each have their own reasons.


When multiple leaders have reasonable claims regarding the same jurisdiction, organizations typically adopt one of the following four coping strategies:

  • Leave it to the most capable leader

  • Leave it to the one with the loudest voice

  • Hand it over to the committee for handling

  • Hand it over to the current budget manager


However, these methods are not applicable to autonomous systemsAIThis matter is of great significance and cannot be resolved solely through political means; The rapid development of technology makes it difficult for committees composed of non experts with conflicting interests to keep up with the pace; And its scope is too broad for any single functional department's existing willingness or ability to cope.


We need a different approach. This method should start with changing the problem, starting from“Who owns itAI?”Transform into“Who owns what andAIRelated decision-making powerThe first problem frames the issue as a zero sum game and triggers enclosure movements, but the second problem decomposes the problem into specific decisions and assigns each decision to the person best suited to make that decision.


Think about governing oneAIAll the choices required by the intelligent agent. Someone needs to decide what the goals of the intelligent agent are and how to measure success. Someone must ensure its construction is secure and integrated into existing systems. Someone must establish rules that specify what data intelligent agents can access and how to make inferences based on this data. Someone must determine to what extent autonomous action is acceptable and what requires manual approval. Someone must figure out how intelligent agents and humans work together, and how introducing intelligent agents can change human work. Someone must also control the initial investment and reasoning budget, and measure whether the investment has yielded returns.


These are different decisions that require different professional knowledge. The Chief Operating Officer or Head of the Profit Center naturally has the authority to decide what goals the intelligent agent should achieve; The Chief Information Officer is responsible for how to securely build and operate them; The Chief Data Officer is responsible for data access and model governance; The Chief Risk Officer and General Counsel are responsible for risk thresholds and escalation mechanisms; The Chief Human Resources Officer is responsible for how humans and intelligent agents share responsibilities; The Chief Financial Officer is responsible for expenses and return on investment. These are not surprising. This is exactly what every executive advocated at that insurance company meeting.


The effect of redefining the problem is that it shifts the discussion from full ownership to creating refined and clear decision-making power. The Chief Information Officer said“I have a technology stackIt's one thing, specifically stating that the Chief Information Officer has the right to choose the underlying model and allows intelligent agents to access the company's record systemAPIIntegration standards, as well as delays in intelligent agent inference interference systems“Emergency stop switch "Control and other related matters are another matter.


In order to shift from ownership to decision-making power, the company will need to establish a decision-making and responsibility allocation diagram. This requires identifying key decisions and assigning each decision to specific executives. To suppress jurisdictional friction, clear decision-making power and clear allocation of responsibilities are needed. But the effectiveness of the allocation chart depends on the process of maintaining it, which means that someone must be responsible for coordinating the work at the coordination level.



The role positioning of Chief AI Officer


This brings us back to the chiefAIOfficer. Many companies establish this position out of confusion because they realize thatAIOwnership cannot belong solely to one person, or because no one in the current executive team has leadership autonomyAIProfessional knowledge or effort in the application process. However, similarly, a new role can only achieve success when it has a reasonable jurisdiction. One“haveAI”The ChiefAIOfficials will face difficulties becauseAITouch everything.


chiefAIOfficials can be responsible for coordinating work at the coordination level. In fact, this means having the decision allocation diagram itself: creating it, maintaining it, identifying gaps, and promoting clarity when the responsibilities assigned to functional owners are unclear. In short, this means delineating appropriate jurisdictional boundaries for other functional leaders. It also means that when new use cases cannot be clearly assigned to existing assignments, relevant executives are convened for coordination——I expect to be in autonomous modeAIIn the early stages, this will occur frequently.


This is a truly innovative jurisdiction. Previously, no current executive had been responsible for cross functional coordination of autonomous systems, but with the autonomous approachAIBecoming a new connecting link in enterprise level organizations has become crucial.



The inevitability of instability


These steps will not eliminate jurisdictional competition, but they can guide it. If the discussion focuses on“Who has which decision-making powerThis will be a healthy conversation, as it will shift from abstraction to a focus on implementation; It will be closer to the needs of the organization; And it will form a framework that can be re examined as the situation changes.


It should be emphasized about who should have whatAIThe issue of decision-making power has no eternal answer. The technology scope is too broad, the impact is too great, and the development is too fast. No solution can remain unchanged.At least in the next few years, it is expected that there will continue to be jurisdictional competition, frequent changes in decision-making power, and with the arrival of a new wave of capabilities, the newly formed balance will be almost equally quickly disrupted.


The insurance company that convened a fruitless meetingCEOThere are several options available. He can wait for the competitors' actions to clarify the direction, or let the person with the loudest voice win. Alternatively, he could acknowledge that every executive in the room is right, and his job is not to pick winners, but to lead the establishment of a coordination system where each of them has the decision-making power that best suits them. This is the core leadership that guides the organization through such a profound and extensive storm of jurisdictional restructuring. That is to design a plan that links ownership with implementation“Map "And then re-examine it every time the technological development exceeds the current version of the organizational structure. In the foreseeable future, this will happen frequently.


Previous: Simplify the procedure for returning illegally obtained cultural relics, France's legislation goes further Next article: The World Bank estimates the issuance of a new AUD 2 billion sustainable development bond with a maturity of 5.5 years

Recommended News

Contact Us

  • Phone: 0017194050155
  • Fax: 0017194050155
  • 邮箱:gelcc1@outlook.com
  • Mobile phone: 0017194050155
  • Address: Room 500, 1234 International Avenue, Washington, D.C. 20001, USA