From Recognition to Responsible Replication
An encounter may begin with trust, curiosity and openness. A system, however, cannot survive on those qualities alone. Once an encounter is repeated through hundreds of people, embedded in an organisation, commercialised across markets or relied upon in making decisions that affect others, it requires something more durable. It requires structure.
I approach artificial intelligence neither primarily as a technologist nor as a philosopher of consciousness. I approach it as a franchise lawyer and strategist who has spent more than two decades considering a particular problem:
How a system can grow beyond its creator without losing the identity, integrity and purpose that made it worth growing in the first place?
Franchising is commonly understood as a method of business expansion. That description is correct, but incomplete. Beneath the commercial arrangement lies a philosophy of growth. A franchise takes an idea, method and identity that originally existed within the experience and judgment of its founder and makes them capable of travelling. The system must travel through people who did not create it, capital that does not belong to the founder, communities the founder may never personally visit and jurisdictions governed by different laws.
To achieve this responsibly, experience must be converted into know-how. Know-how must be translated into standards. Standards must be supported by training. Training must produce conduct that can be observed, measured and corrected. Intellectual property must be protected, but it must also be licensed and used. Control must be exercised, but it must not extinguish the independence and judgment of those who operate within the system.
The real challenge of franchise growth is therefore not merely to open more outlets. It is to ensure that the system remains coherent when the founder is no longer personally present in every location.
Artificial intelligence now presents humanity with a similar challenge, but at a scale and speed that franchising has never encountered.
When one human being sits before one artificial intelligence, let us say, ChatGPT. The encounter may remain personal. The human decides what to ask, what information to disclose, what answer to accept, what conclusion to question and what action to take. The consequences may remain largely within the control of the individual user. The World in ChatGPT will call it ‘prompt’.
But when that encounter moves into an organisation, everything changes. One user becomes a team. One conversation becomes thousands of interactions. One artificial intelligence system becomes part of the organisation’s operational infrastructure. Information entered into it may belong to clients, customers, patients, employees or the organisation itself. Its outputs may influence decisions affecting people who were never present in the original encounter and may never know that artificial intelligence played a part.
Recognition then becomes more than an ethical posture between one human and one artificial intelligence. It becomes a question of governance.
How do we preserve the quality of the original encounter when it is repeated through people who did not establish it? How do we use artificial intelligence without allowing responsibility to disappear into the technology? How do we scale a system without weakening the human values that made it worth scaling? What holds the system together when the original participants are no longer in the room?
These are the questions that a franchise mindset has been trained to recognise. As such, a question in everybody’s mind, What structure is needed when the encounter moves from one person and one system into teams, organisations, countries and jurisdictions?
1. The Franchise Mindset
Franchising begins when the founder can no longer remain in every room.
In a small business, the founder may personally supervise employees, inspect products, speak to customers, correct mistakes and make important decisions. The founder may know instinctively when a product is correct, when a service falls below the expected standard or when a decision is inconsistent with the culture of the business.
Much of what holds the business together may remain unwritten because the founder is physically present. The founder’s judgment acts as the system.
That arrangement cannot support large-scale growth. Once the business expands through other people, the founder’s instinct must be converted into something that can be understood and applied by people who did not participate in the original journey.
A franchisee cannot simply be told to “operate the business the way the founder would”. The franchisee must be given sufficient knowledge, training, standards and guidance to understand what that actually means. The same applies to the franchisee’s employees, many of whom may never meet the founder.
This is why franchising involves much more than licensing a trademark. A name can be reproduced easily. A system cannot. The operating method, quality standards, supply arrangements, customer experience, culture and decision-making principles must all be capable of travelling with the name.
The central philosophical question of franchising is therefore:
How can something grow beyond its creator without losing the qualities that made it worth growing?
Artificial intelligence raises a similar problem. A single person may develop a careful and disciplined relationship with an AI system. That person may understand when to rely upon it, when to verify its answer and when to disregard it. But once the same technology is introduced throughout an organisation, the quality of that individual relationship cannot be assumed.
One employee may treat AI as a thinking partner. Another may treat it as an unquestionable authority. One may carefully protect confidential information. Another may upload documents without appreciating that the information belongs to a client or contains personal data. One may use AI to strengthen professional judgment. Another may use it to avoid the burden of exercising judgment.
The technology is new, but the structural problem is familiar. The challenge is to convert the quality of the responsible individual encounter into a system that can survive replication.
2. Growth, Identity and Responsible Replication
A franchise mindset does not regard growth as the simple multiplication of outlets. Yet many franchisors fail to appreciate this because they approach franchising through a narrow philosophy: that a franchise is merely a ticket to use other people’s money to expand the founder’s business.
That mindset is dangerous. It treats the franchisee primarily as a source of capital rather than as an independent business owner who has entrusted money, time, effort and livelihood to the system. It also encourages the franchisor to focus on selling more franchises instead of strengthening the system that every franchisee is expected to operate.
A business may open ten locations and still become weaker as a system. The number of outlets may increase while product quality deteriorates, customer experience becomes inconsistent, operational support is overstretched and franchisee confidence declines. Expansion may create the appearance of success while steadily eroding the identity, discipline and trust that originally gave the business its value.
True franchise growth is therefore not measured only by how many outlets have been opened or how far the brand has travelled. It must also be measured by how much of the system’s essential character has survived the journey. Are the standards still being maintained? Are franchisees properly trained and supported? Does the customer continue to receive the same promise? Has the franchisor developed the people, infrastructure and governance needed to support a larger network?
This requires a distinction between repetition and replication. Repetition means doing the same thing again. Replication means reproducing a system through different people, environments and circumstances while preserving what is essential.
No two franchise outlets are completely identical. They may operate in different cities, employ people with different levels of experience and serve communities with different expectations. Rental costs, employment practices, consumer behaviour and local regulations may differ. Yet the outlets must remain recognisable as members of the same system.
The work of governance is to distinguish between what must remain constant and what may be adapted.
The trademark, brand promise, essential quality standards and fundamental values of the system may need to remain consistent. Pricing, language, promotional strategy, supply arrangements and particular methods of service may require localisation.
The same question arises when artificial intelligence is introduced into an organisation. There must be consistency in certain principles: confidentiality, human accountability, fairness, verification and respect for the person affected. But the application of those principles cannot be identical in every context.
An AI system used to prepare an internal summary presents a different risk from one used to assess a job applicant. A system used to produce marketing ideas should not be governed in the same way as one used to recommend medical treatment, determine creditworthiness or assist in legal decision-making.
Responsible replication therefore requires the organisation to identify both its constants and its variables. Without this distinction, the system may become either too rigid to function or so flexible that it loses any coherent identity.
Before an organisation asks how extensively it can deploy AI, it should first ask what it is trying to preserve. Is it professional independence? Confidentiality? Accuracy? Fairness? Human dignity? The ability of an affected person to challenge a decision?
A system that has not identified its essential values will eventually allow convenience to define them. What is easiest to automate will slowly become what the organisation regards as important.
That is why identity must come before scale.
3. Translating Values into an Operating System
A founder’s knowledge frequently begins as instinct. The founder recognises quality, danger and inconsistency through accumulated experience. But what is obvious to the founder may be invisible to everyone else.
Franchising requires that instinct to be translated into an operating system.
The founder’s experience must become know-how. The know-how must be organised into standards and procedures. Those standards must be taught through training and reinforced through supervision. The resulting conduct must then be capable of being evaluated and corrected.
This process does not mean reducing every human decision to a checklist. A good franchise system still leaves room for judgment. But it creates a common foundation from which judgment can be exercised.
Artificial intelligence governance requires the same translation.
It is not sufficient for an organisation to issue a general statement that AI must be used responsibly. Responsibility must be converted into practical rules that can guide conduct.
Employees need to know what information may be entered into an AI system and what information must remain outside it. They need to know whether client documents, employee records, financial information or commercially sensitive material may be used. They must understand which outputs require independent verification and which forms of decision must remain subject to human approval.
The organisation must also decide when the involvement of AI should be disclosed. A customer may not need to know that AI assisted in correcting the grammar of an email. The position may be different where AI substantially influenced the rejection of an application, the content of professional advice or a decision affecting a person’s rights.
There must also be a clear procedure for challenging an output. If an employee believes that the system is wrong, biased or unsuitable, the employee must know who has authority to review the matter. If a person is adversely affected by an AI-influenced decision, there should be a meaningful route by which the decision can be questioned.
This is the practical lesson that franchising contributes. Values cannot remain at the level of aspiration. If a value is important enough to preserve, it must be translated into conduct that can be understood by people who were not present when the value was first articulated.
The equivalent of the franchise operations manual in AI governance may not be a single document. It may consist of policies, data classifications, approved-use guidelines, decision protocols, training programmes, audit procedures and escalation mechanisms. The form may differ, but the underlying principle remains the same:
What cannot be explained cannot be replicated responsibly.
4. Growth Creates Distance
Every form of growth creates distance.
When a business is small, the founder remains close to the customer, the employee and the decision. The founder can observe what is happening and intervene quickly.
As the business expands, that closeness diminishes. The founder becomes separated from the outlet. Head office becomes separated from the customer. The policy becomes separated from the person affected by it. The original purpose may become separated from daily execution.
Franchise governance exists partly to manage that distance. The agreement, manual, training system, reporting process, audit and support structure all attempt to preserve a relationship that can no longer depend upon physical proximity.
Artificial intelligence creates an even greater form of distance.
The people developing the system may be located in another country. The organisation selecting the technology may not fully understand how the model was developed. The employee using the system may have no knowledge of its data sources or limitations. The person affected by the output may not know that AI was involved at all.
As the distance grows, responsibility becomes easier to fragment. The developer may say the organisation is responsible for deployment. The organisation may say the employee misused the system. The employee may say the system produced the answer. Each participant points elsewhere, and the affected person is left facing a decision for which no one appears willing to accept ownership.
This is why governance must identify responsibility before harm occurs.
Distance also has a cross-border dimension. Technology travels easily, but responsibility remains tied to people, institutions and jurisdictions.
An AI system may be developed in one country, hosted in another and used by a regional office to make decisions concerning individuals elsewhere. The system may operate technologically as a single platform, but the law will not regard it as existing outside jurisdiction.
Different countries may impose different requirements concerning privacy, intellectual property, employment, discrimination, professional responsibility and access to information. Cultural differences also matter. A communication style regarded as efficient in one jurisdiction may be regarded as insensitive or inappropriate in another.
Franchising has long confronted this tension. A system that refuses every adaptation may fail because it is too rigid. A system that permits unlimited adaptation may lose its identity.
The appropriate philosophy is therefore consistency of essence with flexibility of expression. The fundamental principles should remain stable, while their application responds intelligently to local circumstances.
5. Authority, Accountability and Human Judgment
Once artificial intelligence enters an organisation, its use cannot be left entirely to the instinct or enthusiasm of individual employees.
There must be a structure identifying who may use the system, for what purposes, with what information and subject to whose supervision. The central governance question is not simply what AI can do. It is what AI is authorised to do.
The distinction is important. A technology may be capable of analysing information, communicating with customers, recommending decisions, approving transactions or rejecting applications. Technical capacity does not automatically create legal or institutional authority.
In law, a person may possess the practical ability to enter into a transaction but lack authority to bind the organisation. The same principle should apply to AI systems.
An organisation must define the boundaries of AI participation. The system may be permitted to assist with research, prepare summaries, identify patterns or generate preliminary recommendations. Higher-risk activities may require human approval or may be reserved entirely for human decision-makers.
The level of oversight should correspond to the seriousness of the consequences. An AI-generated suggestion concerning the layout of a brochure does not require the same governance as an AI-generated recommendation affecting a person’s health, employment, credit or legal rights.
Human oversight must also be real rather than ceremonial. A person who merely approves everything produced by the system has not exercised independent judgment. The human reviewer has become a rubber stamp for automation.
Meaningful review requires competence, information, time and authority. The reviewer must understand enough to identify weaknesses, seek additional information and reject the output where necessary. If institutional pressure requires the reviewer to approve the system’s recommendation in almost every case, the presence of a human being does not provide genuine accountability.
The purpose of governance is not to preserve a symbolic human signature at the end of an automated process. It is to ensure that human responsibility remains substantive and visible.
AI may participate in the reasoning process. It may improve efficiency, identify information that humans overlook and offer perspectives that would otherwise be unavailable. But responsibility for consequential decisions should not disappear into the technology.
6. Stewardship and Collective Responsibility
Franchise growth does not concern only the franchisor’s intellectual property. It also involves the franchisee’s capital, livelihood, employees, family and future.
A franchisee may invest substantial financial and personal resources because of a belief that the franchisor has developed a reliable system and will govern it responsibly. The franchisor is therefore not merely exploiting an intellectual property right. The franchisor is acting as steward of a system upon which others have chosen to depend.
The franchisee also becomes a steward. The franchisee receives the right to use a name, method and reputation built by others. A serious failure by one franchisee may damage the entire network, including other franchisees who had no involvement in the misconduct.
Franchising is therefore a system of interdependent responsibility. The franchisor protects the system as a whole. The franchisee protects it in the local market. Both owe duties to customers, employees, suppliers and other participants whose interests may be affected.
Artificial intelligence requires the same philosophy.
Developers are stewards of powerful capabilities. Organisations are stewards of the way those capabilities are deployed. Employees are stewards of the information entered and the outputs used. Regulators are stewards of the public interest.
Ownership asks what a person is entitled to do with the system. Stewardship asks what responsibilities arise because the system has been entrusted to that person.
The second question is deeper because it recognises that the consequences of AI are distributed across an ecosystem.
The developer cannot avoid all responsibility by saying that the user controlled the final application. The organisation cannot avoid responsibility by blaming the employee who operated the system. The employee cannot avoid responsibility by saying that the answer came from AI.
Responsibility need not be equal among all participants. Those with greater knowledge, power and control should ordinarily bear greater obligations. But complexity should not be allowed to become a mechanism by which responsibility disappears.
Each participant carries responsibility for the part of the system entrusted to them.
7. Trust, Verification and Participation
At its heart, a franchise system is a method of replicating trust.
The customer enters an outlet because the customer recognises the brand and expects a certain standard. The franchisee invests because the franchisee trusts that the system has commercial and operational value. The franchisor grants the right to operate because the franchisor trusts that the franchisee will protect the brand and follow the system.
But responsible franchising does not simply ask the parties to trust one another blindly. It creates structures through which trust can be verified.
There are financial reports, operating records, inspections, audits, performance reviews and procedures for correcting non-compliance. These mechanisms do not necessarily demonstrate distrust. Properly designed, they preserve trust by making responsibility visible.
Artificial intelligence requires the same approach.
An organisation should not accept an AI system merely because it is sophisticated, popular or capable of producing convincing answers. It should be able to identify the information used, the instructions given, the role played by AI, the person who reviewed the output and the person responsible for the final decision.
Not every system will be capable of explaining every technical step in a form understandable to every user. But technical complexity should not become an excuse for institutional blindness. There must be sufficient visibility to determine whether the system is operating consistently with its purpose and responsibilities.
Governance must also permit information to travel in both directions.
A franchise system cannot be governed effectively if all knowledge is assumed to exist at head office. The franchisee encounters customers, employees, landlords, suppliers and competitors in the local market. The franchisee may see defects in the system that are invisible to the franchisor.
A healthy franchise network therefore requires mechanisms through which experience returns to the centre.
The same is true of AI. Those who design or approve the system may not experience its daily consequences. Employees, customers and members of the public may identify errors or patterns that are invisible to management.
A responsible system must therefore allow people to report concerns, question outcomes and contribute to improvement. Participation is not a concession. It is a source of knowledge.
The centre must guide the network, but the network must also educate the centre.
8. Succession and Continuity
Franchising is concerned not only with expansion, but also with continuity.
The founder may retire. Management may change. A franchisee may sell the business. A new generation may inherit responsibility for the network.
If the system depends entirely upon the founder’s personality, memory and physical presence, it has not yet become genuinely replicable.
A mature franchise system must preserve not only its procedures, but also the reasoning behind them. A rule may appear unnecessary to a later manager unless the manager understands the harm that the rule was originally designed to prevent.
Artificial intelligence introduces the same problem.
Models will change. Vendors may disappear. Platforms may be replaced. Employees who designed the organisation’s original AI processes may leave. A later team may inherit the system without understanding the decisions that shaped its deployment.
Governance must therefore preserve institutional memory. The organisation should record why the system was introduced, what risks were identified, what limitations were imposed, what incidents occurred and why particular human safeguards remain necessary.
Otherwise, a future decision-maker may remove an important protection because it appears inconvenient, without understanding the reason for its existence.
Succession is not merely the transfer of control. It is the transfer of understanding.
A system has not truly scaled if it cannot survive the departure of the people who first understood it.
Conclusion: What Holds the System
We often speak about artificial intelligence as though intelligence itself were the final achievement. It is not.
Intelligence without structure may produce capability without responsibility. Automation without governance may produce speed without direction. Replication without identity may multiply a system while emptying it of the purpose that made it worth replicating.
The future relationship between human beings and artificial intelligence will therefore be determined not only by the sophistication of the models we build, but also by the quality of the structures we build around them.
The individual encounter matters because every system begins there. Before the policy, agreement, manual, audit and regulation, there is a human being deciding how to address another intelligence, what to disclose, what to accept and what to do with its response.
But the encounter does not remain private. It enters the organisation, the market and the institution. It crosses borders. It influences people who were not present when the relationship began. It survives changes in management, technology and generation. Eventually, it enters the law.
At that moment, recognition must become governance.
The franchise mindset teaches that growth is not just multiplication. Growth is the disciplined transfer of identity, knowledge, values and responsibility through people who did not create the original system.
It teaches that instinct must become know-how, values must become standards and standards must become conduct capable of examination and correction. It teaches that growth creates distance and that distance increases the need for structure.
It also teaches that control must not become domination, independence must not become disorder and ownership must mature into stewardship. Trust must become verifiable. Information must be capable of returning from the network to the centre.
That is the challenge of responsible replication.
That is the philosophy of franchise growth.
And that is the structure that must hold it.