Dr. Ahmed Mokhtar: The experiences that most shaped my understanding of law came from moving repeatedly between theory, institutions, and real disputes. My academic background in public law trained me to think comparatively and to look beyond the text of a rule to the constitutional principles, institutional design, and public interests behind it. My doctoral and later research work, particularly across Egypt, France, and the UAE, reinforced the idea that the same legal principle can operate very differently depending on the strength of institutions, administrative culture, and judicial control. Legal practice added a different dimension: working on construction contracts, commercial transactions, disputes, litigation strategy, arbitration-related matters, and regulatory compliance taught me that law is not tested in the abstract. It is tested under pressure, when wording, evidence, timing, notices, and risk allocation can determine the outcome of a dispute. My work in legal education and academic quality also shaped me significantly. Teaching and supervising students made me more conscious of the gap that can exist between knowing legal rules and being able to reason with them in practice. More recently, my research on artificial intelligence, digital governance, digital identity, and automated public decision-making has changed the questions I ask about law. I now see law not only as a system of rules, but as a framework for controlling power, allocating responsibility, protecting rights, and preserving human judgment when technology increasingly influences public and private decisions. So today, I understand law as something living: it must be principled, but also workable. It must provide certainty, but remain capable of responding to new forms of power, risk, and technology.

Dr. Ahmed Mokhtar: Working across Egypt and the Gulf changed my understanding of legal effectiveness in a very practical way. I came to see that a legal system is not effective simply because it has detailed legislation or sophisticated institutions on paper. Its real strength appears in how consistently rules are implemented, how quickly rights can be enforced, and how clearly public and private actors understand their obligations. My experience in academia, public law, and corporate legal practice exposed me to different administrative cultures, regulatory environments, and approaches to dispute resolution. In Egypt, I developed a strong appreciation for doctrinal depth, administrative law traditions, and judicial reasoning. In the Gulf, I became more attentive to speed of implementation, regulatory modernization, institutional efficiency, and the close relationship between law and economic development. My work in Kuwait and Saudi Arabia, particularly in contracts, compliance, disputes, and institutional governance, reinforced this perspective. What I have learned is that an effective legal system needs four elements working together: clear rules, capable institutions, predictable enforcement, and accessible remedies. If one of these is missing, legal certainty weakens. I also believe effectiveness requires adaptability. A system that cannot respond to digital transformation, artificial intelligence, new forms of contracting, and cross-border commerce will gradually become less relevant, even if its traditional legal framework is strong. So today, I judge legal effectiveness not by the volume of legislation, but by whether the system produces trust. Citizens, investors, contractors, and public authorities should be able to predict how the law will operate, what procedures they must follow, and what remedies are available when something goes wrong.

Dr. Ahmed Mokhtar: Technology became a legal question for me when I began to see that digital systems were no longer merely supporting public administration, but were increasingly influencing how public power is exercised. My research on artificial intelligence, digital identity, data governance, predictive government, and automated decision-making made this shift very clear. Once an algorithm helps determine access to a public service, assesses a person’s risk, influences an administrative decision, or structures how government data is used, the issue is no longer simply technical. It becomes a question of legality, accountability, transparency, equality, and fundamental rights. A major turning point for me was studying administrative liability for damage caused by AI systems in public services. That work forced me to confront a practical question: if an automated system causes harm, can the administration simply say that “the system” made the decision? My answer is no. Public authority cannot outsource legal responsibility to technology. The more automation enters decision-making, the more important it becomes to preserve traceability, human oversight, and the ability to challenge the outcome. I also became increasingly interested in digital identity and predictive government because these areas show how technology can affect a person before any traditional legal dispute even begins. Data can classify, rank, predict, and sometimes exclude. That creates a new form of public power. So, for me, the key moment was when technology ceased to be merely an administrative tool and became part of the decision itself. At that point, lawyers must ask not only whether the technology works, but whether its use is lawful, explainable, proportionate, reviewable, and consistent with constitutional rights.

Dr. Ahmed Mokhtar: Legal practice taught me that the law is often decided in the details that academic study tends to treat as secondary: timing, evidence, drafting, procedure, and the sequence in which decisions are made. In academic work, we naturally focus on legal principles, judicial reasoning, and doctrinal coherence. Practice showed me that a strong legal position can still be weakened by a missed notice, an ambiguous clause, poor documentation, or a failure to preserve evidence. This became especially clear in construction and commercial matters, where delay claims, variations, payment disputes, contractual notices, and risk allocation can determine the entire trajectory of a dispute. My current work in legal affairs has made that practical dimension impossible to ignore. At the same time, scholarship later taught me something equally important about practice: immediate solutions are not always the best solutions. Academic research encourages distance, comparison, and structure. It asks whether a practical response is consistent with broader principles, whether a recurring dispute reveals a deeper drafting or governance problem, and whether another legal system has already developed a more effective approach. That relationship between practice and scholarship has become central to the way I work. Practice gives me the facts, pressures, and consequences. Scholarship gives me the framework to understand patterns and avoid treating every problem as isolated. I now see the two as mutually corrective. Practice prevents legal thinking from becoming abstract. Scholarship prevents practice from becoming merely reactive. The strongest legal judgment, in my view, comes from combining both: understanding what the law means in principle, and how it actually behaves when tested by real institutions, real contracts, and real disputes.

Dr. Ahmed Mokhtar: When an algorithm begins to influence an administrative decision, the legal problem changes at several levels at once. The key issue is no longer only whether the final decision is lawful, but also how that decision was produced, what data shaped it, whether the logic can be understood, and who remains accountable for the result. Traditional administrative law assumes a human decision-maker whose reasoning can be examined through competence, procedure, purpose, proportionality, equality, and judicial review. Algorithmic systems complicate each of these elements. They may rely on large datasets, probabilistic assessments, automated classifications, or models that even the public authority itself cannot fully explain. That creates a risk that discretion does not disappear, but is simply hidden inside data selection, model design, thresholds, and technical parameters. This is why I believe algorithmic administration requires an expanded concept of legality. It should include not only lawful authority and proper procedure, but also data quality, traceability, explainability, meaningful human oversight, and the possibility of effective challenge. My research on AI in public administration has focused precisely on these concerns, particularly legality, transparency, liability, and the constitutional protection of digital rights. A practical example is an automated system used to assess eligibility for a public benefit or to identify administrative risk. Even if the system is statistically accurate, a citizen should not be adversely affected by a process that is opaque, based on incorrect data, or impossible to contest. So the central change is this: administrative law must move from reviewing only the decision to reviewing the decision-making architecture itself. Technology may assist public authorities, but it should never create an accountability gap.

Dr. Ahmed Mokhtar: In my view, accountability should not be assigned to a single actor automatically. It should follow the actual distribution of control, knowledge, and responsibility across the system. The public authority must remain the primary point of legal accountability because it is the entity exercising public power and making, adopting, or relying on the decision. It should not be able to avoid responsibility simply because an AI system was developed or supplied by a private company. This is especially important where the decision affects rights, public benefits, licenses, sanctions, or access to services. At the same time, responsibility may also extend to the developer, vendor, or operator where the harm results from defective design, poor training data, inadequate testing, misleading performance claims, negligent implementation, or failure to identify known risks. In complex systems, liability may therefore need to be shared. My own research on administrative liability for harm caused by AI systems in public services has reinforced this point. The legal question should not be “Who touched the system last?” but rather: Who designed it? Who selected the data? Who configured the thresholds? Who approved its use? Who had the ability to intervene? And who benefited from or exercised the resulting public power? A practical framework should combine several layers of accountability: institutional responsibility of the public authority, contractual and professional liability of technology providers, operational responsibility of those deploying the system, and effective judicial remedies for the affected person. The central principle, however, should remain clear: automation must never become a legal shield. Where public power is exercised through AI, responsibility must remain traceable, reviewable, and enforceable.

Dr. Ahmed Mokhtar: Meaningful human oversight should be more than a formal “human-in-the-loop” requirement. In practice, the human reviewer must have real authority, sufficient information, enough time, and the institutional freedom to question or override the system. For me, meaningful oversight begins before the decision is made. Public authorities should define which decisions may be automated, which require mandatory human review, and which should never be delegated to an automated system because of their impact on rights or legal status. This is particularly important in areas involving sanctions, access to public benefits, licensing, disciplinary measures, or decisions that may seriously affect an individual. The reviewer must also understand the basis of the system’s output. That does not mean every official must be a data scientist, but the authority should be able to explain the relevant data, factors, confidence level, limitations, and known risks. A person who merely clicks “approve” without understanding or questioning the recommendation is not exercising oversight. My research on AI-driven governance, administrative liability, digital rights, and automated public decision-making has reinforced my view that human oversight must be connected to legality, transparency, and accountability, not treated as a procedural formality. In practical terms, I would require at least five safeguards: clear responsibility for the final decision, access to the relevant reasoning and data, authority to reject the automated recommendation, documented reasons when the system is followed or overridden, and an effective channel for the affected person to request human reconsideration. The test is simple: if the human reviewer cannot realistically change the outcome, then the oversight is not meaningful. It is only symbolic.

Dr. Ahmed Mokhtar: A court should begin from a simple but important premise: technical accuracy does not automatically produce legal fairness. An AI system may be statistically reliable and still generate outcomes that are discriminatory, disproportionate, opaque, or inconsistent with fundamental rights. For that reason, judicial review should not stop at asking whether the system “worked correctly” in technical terms. The court should also examine whether the decision respected legality, equality, due process, proportionality, and the right to an effective remedy. In practice, I would distinguish three levels of review. First, the court should verify the legal basis for using the system and whether the public authority acted within its powers. Second, it should examine the process: the quality of the data, the possibility of bias, the transparency of the relevant factors, and whether meaningful human oversight existed. Third, it should assess the substantive effect of the decision on the individual. This is where fairness becomes essential. A decision may be efficient for the administration but still impose an excessive or unjustified burden on a person. In such a case, efficiency cannot override legality or proportionality. My comparative work on AI, digital rights, and public administration has strengthened my view that courts should treat algorithmic systems as part of the administrative decision-making process, not as neutral technical tools outside legal scrutiny. So if a technically accurate AI decision produces an unfair result, the court should ask a deeper question: not only whether the model was correct, but whether the use of that model was legally justified and constitutionally acceptable in the circumstances.

Dr. Ahmed Mokhtar: Governments do not need to explain every technical detail of an AI system, but they must explain enough for a citizen to understand why a decision affected them and how they can challenge it. For me, a meaningful explanation should answer four practical questions: What was decided? What were the main factors that influenced the decision? What data about the individual was used? And what can the person do if the result is wrong or unfair? This is especially important in public administration because the citizen is not dealing with a private recommendation engine; they are dealing with an exercise of public power. If an AI-assisted system contributes to denying a license, restricting access to a public service, prioritizing an inspection, or assessing administrative risk, the authority should be able to translate the technical process into legally understandable reasons. The explanation does not need to disclose source code or reveal protected technical secrets. It should instead provide the relevant decision logic, the decisive data points, known limitations, and the role played by human review. My work on digital identity, data governance, predictive government, and AI-supported public decision-making has reinforced the importance of this distinction between technical complexity and legal intelligibility. I would also add one safeguard: the explanation must be useful for challenge, not merely informative. If a citizen cannot identify an error, contest the data, or request reconsideration, then the explanation is incomplete. So the real standard should not be “Can the government explain the algorithm?” but rather: “Can the government explain the decision in a way that allows the citizen to understand it, contest it, and obtain meaningful review?”

Dr. Ahmed Mokhtar: Yes. An AI system can be technically reliable and still be constitutionally unacceptable. Technical reliability usually asks whether the system performs consistently, accurately, and within its intended parameters. Constitutional review asks a different set of questions: whether the system respects legality, equality, due process, privacy, proportionality, and the effective protection of fundamental rights. A system may therefore be highly accurate and still be unlawful if it relies on discriminatory data, produces unjustified differential treatment, interferes excessively with privacy, or prevents an affected person from understanding and challenging the decision. In my view, courts should examine at least five elements. First, the legal basis for using the system and whether the public authority acted within its competence. Second, the quality and relevance of the data used. Third, whether the system creates direct or indirect discrimination. Fourth, whether the decision is proportionate to the legitimate public objective pursued. Fifth, whether the individual has access to an effective remedy, including meaningful reasons and human review. This concern is central to my research on AI, digital rights, digital identity, and predictive government, where the core issue is not whether technology is efficient, but whether its use remains compatible with constitutional guarantees. A predictive system illustrates the problem well. It may identify risk with impressive statistical accuracy, but if that prediction leads to adverse treatment before any actual wrongdoing, the constitutional issue becomes serious. Accuracy cannot by itself justify preventive restrictions on rights. So courts should look beyond performance metrics. The decisive question is whether the use of the system preserves the constitutional relationship between the individual and public power. Efficiency is valuable, but constitutional legitimacy must remain the higher standard.

Dr. Ahmed Mokhtar: Digital identity should never be treated merely as a technical credential. Once it becomes the gateway to public services, benefits, records, payments, or interaction with government, it becomes part of the legal relationship between the individual and the state. My view is that citizens should receive a bundle of protections alongside any digital identity system. First, there must be a clear legal basis defining what data may be collected, for what purposes, and for how long. Second, governments should apply data minimisation: a digital identity system should not become a justification for collecting information that is unnecessary for the service concerned. Third, individuals should have practical rights to access, correct, and where legally appropriate challenge the use of inaccurate or outdated data. Security is equally important. Identity systems concentrate sensitive personal information, so strong cybersecurity, authentication, breach-response mechanisms, and institutional accountability are essential. But protection must go beyond cybersecurity. The system should also guard against profiling, function creep, discrimination, and the use of identity data for purposes unrelated to the original legal objective. My research on the legal framework of digital identity in Egypt, the UAE, France, and the European Union has reinforced this concern: digital identity is not only about proving who a person is; it is also about controlling who may know what about that person, for what purpose, and with what legal consequences. I would add one further principle: no citizen should lose meaningful access to essential public services simply because a digital identity system fails, excludes them, or misidentifies them. There must always be correction procedures, human assistance, and effective legal remedies. A trustworthy digital identity system should therefore achieve two goals at the same time: make public services easier to access, while making the individual more—not less—protected against misuse of state-held data.

Dr. Ahmed Mokhtar: The line should be drawn where the use of data ceases to be necessary for delivering a legitimate public service and begins to create continuous, disproportionate, or insufficiently controlled observation of individuals. My research on digital identity, data governance, predictive government, and digital rights has convinced me that data-driven government can improve public administration significantly, but only if it remains tied to clear legal purposes and effective safeguards. In practical terms, I would apply several tests. First, purpose: the government should collect and use data only for a clearly defined and lawful objective. Second, necessity: authorities should not process more data than is reasonably required to achieve that objective. Third, proportionality: the public benefit gained from data use must justify the degree of intrusion into private life. Fourth, retention and access: governments should define who may access the data, how long it is retained, and whether it may be reused for a different purpose. Fifth, accountability: citizens should know when significant data processing affects them and should have access to correction, review, and legal remedies. A particularly serious concern arises when data collected for one legitimate service is later combined with other databases to create profiles, predictions, or risk scores without a clear legal basis. At that point, useful administration can gradually become surveillance. I therefore do not see the distinction as a choice between innovation and privacy. A mature legal system should be capable of protecting both. The objective should be data-enabled government without creating a surveillance state.

Dr. Ahmed Mokhtar: A prediction of fraud or administrative risk should never be treated as proof. It should be treated as a signal that may justify further review, but not as a sufficient basis for an adverse legal consequence by itself. This distinction is central to my research on predictive government and digital rights. Once public authorities use data to classify individuals according to future risk, the legal concern becomes much greater because a person may be affected not by what they have actually done, but by what a system predicts they might do. Before such a prediction affects a person, several safeguards should exist. First, the data must be relevant, accurate, lawfully obtained, and periodically updated. Second, the model should be tested for bias, especially where historical data may reflect past discrimination or institutional imbalance. Third, the person should not be subject to an adverse decision solely on the basis of an automated risk score. There should be meaningful human review capable of examining context and correcting error. Fourth, the affected person should be informed, at least in substance, that predictive analysis played a role in the decision and should have the opportunity to challenge inaccurate data or misleading assumptions. Fifth, the consequences should be proportionate. A prediction may justify verification, audit, or closer review, but it should not automatically justify sanctions, exclusion, or deprivation of rights. I would also insist on institutional audit trails: authorities should be able to reconstruct how the prediction was produced, what data was used, who approved reliance on it, and what human intervention occurred. The governing principle should be clear: prediction may trigger scrutiny, but it should not replace proof, procedure, or individual justice.

Dr. Ahmed Mokhtar: Yes. In my view, the right to challenge an automated public decision should extend beyond the final outcome to the significant data, assumptions, and reasoning that materially influenced it. If a citizen is denied a public benefit or service, procedural fairness requires more than a simple notification saying that the application was rejected. The person should be able to understand the basis of the decision well enough to identify possible factual errors, incomplete data, inappropriate classifications, or unreasonable assumptions. This is especially important where automated systems rely on data aggregation or risk scoring. A decision may appear objective while in fact being shaped by outdated information, incorrect records, or variables that indirectly disadvantage certain groups. Without access to the important data and reasoning, the right to appeal becomes largely theoretical. I do not mean that every citizen must receive the full source code or proprietary technical documentation. The legal standard should instead focus on material transparency: disclosure of the decisive data, the main factors relied upon, the role of the automated system, and the reasons why those factors led to the adverse outcome. My research on digital identity, data governance, and automated public decision-making has reinforced the importance of this connection between transparency and effective legal protection. There should also be a practical right to correct inaccurate data and to request meaningful human reconsideration. Otherwise, the citizen may be trapped in a process where the same flawed data repeatedly reproduces the same result. So the principle is straightforward: a person cannot effectively challenge a decision if the evidence and logic behind that decision remain inaccessible.

Dr. Ahmed Mokhtar: Europe offers Arab legal systems several valuable lessons, particularly in the areas of institutional accountability, data protection, judicial review, transparency, and the regulation of emerging technologies. But I do not believe that legal development should become a process of copying European models. My comparative work across Egypt, the UAE, France, and the European context has reinforced this point. Different legal systems may share the same concern—such as digital rights, automated decision-making, or administrative accountability—while still requiring different institutional solutions because their constitutional structures, administrative traditions, regulatory capacities, and social priorities are not identical. What Arab legal systems should learn from Europe is mainly methodological. Europe has increasingly emphasized clear legal bases for data use, proportionality, independent oversight, rights of access and challenge, and stronger accountability in the use of automated systems. These are principles worth adopting because they strengthen trust and legal certainty. But the Arab region should develop its own approach in at least three areas. First, digital government in many Gulf countries is advancing very quickly, sometimes faster than traditional legislative cycles. Regulation therefore needs to be flexible enough to support innovation without weakening legal safeguards. Second, Arab legal systems should build rules that reflect their own administrative structures rather than importing institutional models that may not fit. Third, regional cooperation is important because digital platforms, data flows, and AI systems increasingly operate across borders. So I would describe the ideal approach as comparative, not imitative. Arab legal systems should learn from Europe’s experience, including both its strengths and its regulatory difficulties, but then design rules that are legally sound, institutionally realistic, and suited to their own social and economic context.

Dr. Ahmed Mokhtar: Governments may use AI to identify patterns and anticipate risks, but prediction should support public administration, not become a substitute for evidence, due process, or individual responsibility. This issue is central to my research on the constitutional legitimacy of predictive government. The legal danger begins when a prediction moves from being an analytical tool to becoming the basis for an adverse decision against a person before any actual violation has occurred. In my view, predictive systems may legitimately be used for purposes such as allocating inspection resources, detecting unusual patterns, prioritising administrative review, or identifying areas that deserve closer scrutiny. But the more serious the consequence for the individual, the stronger the safeguards must be. A risk score should not by itself justify denial of a benefit, a sanction, exclusion from a service, or any comparable restriction on legal rights. The essential distinction is between prediction as a trigger for examination and prediction as a basis for punishment or deprivation. I would require several safeguards: a clear legal basis, reliable and relevant data, regular testing for bias and error, proportionality, meaningful human review, transparency about the role of the predictive system, and an effective opportunity for the individual to challenge the data and the conclusion drawn from it. There is also a deeper constitutional concern. Predictive government can gradually shift public law from responding to conduct to governing people according to probability. That shift requires great caution because individuals should not be reduced to statistical profiles. So governments should be allowed to anticipate risk, but not to treat prediction as guilt. The closer a predictive system comes to affecting rights, status, or legal opportunities, the more law must insist on evidence, human judgment, and review.

Dr. Ahmed Mokhtar: Yes. A predictive system can become legally dangerous precisely because it is highly efficient. The problem is not accuracy by itself. The problem begins when statistical prediction is converted into differential legal treatment. A system may correctly identify patterns at population level, yet still produce unfair consequences for an individual who is judged not by what they have done, but by what the system predicts they may do. This is one of the central concerns in my work on the constitutional legitimacy of predictive government. The more efficient the system becomes, the greater the temptation for public authorities to rely on it automatically. That creates a risk of what I would call administrative determinism: treating a probability as though it were a fact. For example, a person may be classified as high risk because they resemble a statistical profile associated with fraud, non-compliance, or administrative abuse. Even if the model is highly accurate overall, the legal question remains whether that classification justifies treating that individual differently before any actual violation has occurred. Courts and public authorities should therefore examine not only predictive accuracy, but also the legal consequences attached to the prediction. They should ask whether the differential treatment is based on a legitimate objective, whether it is necessary and proportionate, whether less restrictive alternatives exist, and whether the individual can challenge the data and inference. The deeper danger is that efficiency may normalize unequal treatment before wrongdoing. That would shift public law from responsibility for conduct to governance by probability. So, in my view, prediction may inform administrative attention, but it should never become an automatic basis for disadvantage. The law must preserve the distinction between risk assessment and legal responsibility.

Dr. Ahmed Mokhtar: Ten years from now, citizens should expect an AI-enabled public administration that is faster and more responsive, but also more transparent, reviewable, and accountable than many current systems. The real measure of trust should not be how much AI government uses, but how responsibly it uses it. A trustworthy administration should be able to explain when AI has influenced a decision, identify the data and criteria that mattered, preserve meaningful human oversight, and provide an effective path for review when the citizen believes the result is wrong. This reflects the concerns at the centre of my work on AI-driven governance, digital rights, and public administration. I would expect several practical improvements. Routine services should become quicker and more personalized. Administrative processes should involve less duplication and fewer unnecessary delays. Data should be used to identify public needs earlier, improve resource allocation, and support better policy design. But the same systems should be designed to avoid discrimination, excessive surveillance, and opaque decision-making. Citizens should also have what I would call digital procedural rights: the right to know when AI materially affects them, the right to challenge inaccurate data, the right to meaningful reasons, and the right to human reconsideration in significant cases. My concern is that digital efficiency should not come at the cost of legal dignity. A highly automated administration may still fail if citizens feel they are dealing with an invisible system they cannot question. So, in ten years, the best AI-enabled public administration should feel both technologically advanced and legally human: faster in service, better in prediction, but still governed by legality, fairness, accountability, and respect for individual rights.

Dr. Ahmed Mokhtar: The greater challenge, in my view, is not the technology itself. It is whether institutions, regulators, courts, and lawyers are able to understand it well enough to govern it effectively. Technology will continue to evolve faster than legislation. That means legal systems cannot rely on the traditional model of waiting for every new technical development and then producing a separate rule for it. What institutions need instead is a strong governance framework built around durable principles: legality, accountability, transparency, proportionality, human oversight, data protection, and effective remedies. My work on AI regulation, digital governance, administrative liability, and the evaluation of AI-assisted legal outputs has made this very clear to me. The problem often arises not because the technology is inherently unmanageable, but because institutions lack technical literacy, clear lines of responsibility, or procedures for auditing and challenging automated systems. Lawyers also need to change. It is no longer enough to understand only statutes and cases. Lawyers increasingly need to understand how data is used, how automated systems reach outputs, what risks arise from model design, and how contractual, administrative, constitutional, and regulatory rules interact with technology. This does not mean lawyers must become engineers. It means they must become capable translators between law, technology, institutions, and rights. I therefore see AI governance as an institutional capacity problem as much as a regulatory one. The strongest legal framework will fail if the institutions applying it are weak, undertrained, or unable to question technical systems. So the central challenge is not whether AI can be governed. It can. The real question is whether our institutions and legal professions can develop fast enough to govern it responsibly.

Dr. Ahmed Mokhtar: If I had to establish one non-negotiable principle for government use of AI, it would be this: public authority must remain legally accountable for every AI-assisted decision that affects rights, status, or access to public services. That principle is more important than any particular technical rule because it prevents responsibility from disappearing behind the system. A government may use advanced tools, external vendors, predictive models, or automated platforms, but it should never be able to say that “the algorithm decided” as if that ends the legal inquiry. My research on administrative liability, digital governance, automated decision-making, and digital rights has reinforced this position. The citizen must always be able to identify who is responsible, understand the essential basis of the decision, challenge errors in data or reasoning, and obtain meaningful human and judicial review where necessary. This principle also has a practical consequence: accountability must be designed into the system from the beginning. Public authorities should define who approves the model, who monitors its performance, who can override its output, how decisions are documented, and what remedy is available when harm occurs. In my view, many other safeguards follow from this single rule: transparency, explainability, human oversight, auditability, proportionality, and effective remedies. So if I had to reduce responsible government use of AI to one principle, it would be this: technology may assist the exercise of public power, but it must never obscure or weaken legal responsibility for that power.

Dr. Ahmed Mokhtar: What changes most fundamentally is that part of the contract moves from legal text into technical execution. In a traditional government contract, obligations are interpreted and implemented through human decisions, notices, certificates, approvals, and administrative judgment. Once smart or automated mechanisms are introduced, some of those steps may be triggered automatically by data, software, or predefined conditions. That can improve speed, consistency, and auditability, but it also changes where legal risk sits. My research on smart administrative contracts and digital public procurement has focused on this transition. In my view, at least five issues become critical. First, the contract must clearly identify which obligations are automated and which remain subject to human or administrative judgment. Second, the legal text and the code must be aligned. If the software executes something different from what the contract legally requires, the legal agreement must remain the controlling reference. Third, responsibility must be allocated for technical failure, incorrect data, cybersecurity incidents, and errors in automated execution. Fourth, there must be a mechanism for suspension or human intervention where exceptional circumstances arise. Fifth, the system must preserve an audit trail showing how and why an automated action occurred. This is particularly important in public contracts because public law principles do not disappear merely because performance becomes digital. Transparency, equality, public interest, oversight, and the right to challenge administrative action must still be protected. So, smart mechanisms do not replace contract law. They make drafting, governance, and risk allocation even more important. The technology may automate performance, but it cannot automate away legal judgment.

Dr. Ahmed Mokhtar: Smart contracts can support transparency, equality, public interest, oversight, and the right to challenge, but only if those principles are designed into the legal and technical architecture from the beginning. Technology does not guarantee them automatically. My research on smart administrative contracts and digital public procurement has reinforced this point. A smart contract may increase consistency, reduce manual interference, and create a reliable record of transactions. These are real advantages. But in public contracting, efficiency is not the only value. Public contracts must also remain subject to legality, accountability, equal treatment, public interest considerations, and effective review. The main difficulty is that code operates through predefined conditions, while public law often requires context, interpretation, proportionality, and discretion. A system may execute a contractual consequence exactly as programmed, yet still produce a legally questionable result if an exceptional circumstance, procedural defect, or public-interest consideration has not been properly incorporated. For that reason, I would not support a model in which code becomes final and unchallengeable. There should be clear rules on when automated execution is permitted, when human intervention is mandatory, how exceptional circumstances are handled, and how an affected party may suspend, review, or challenge an automated outcome. Transparency also requires more than publishing the contract. The parties should understand which obligations are automated, what data triggers execution, who controls the system, and how errors are corrected. So my answer is yes, smart contracts can respect public-law principles, but only under one condition: the technology must remain subordinate to the legal framework, not the other way around.

Dr. Ahmed Mokhtar: Digital procurement can reduce arbitrariness, standardize procedures, and create stronger audit trails, but automation can also move discretion into places that are less visible: system design, eligibility criteria, scoring models, data inputs, and technical rules. That is why I do not believe the answer is to choose between automation and human control. The real objective should be accountable automation. My work on smart government contracts and digital public procurement has made me particularly attentive to this issue. In practice, governments should build accountability into the procurement system at several stages. First, the rules embedded in the digital platform should be legally traceable. It should be possible to understand how eligibility, ranking, exclusion, or evaluation criteria operate. Second, authorities should preserve complete audit trails showing what data was used, what action was triggered, and who approved the relevant configuration. Third, high-impact decisions should not be fully insulated from human review. There must be a clear mechanism to challenge erroneous data, technical malfunction, or an automated outcome that produces an unreasonable result. I would also require periodic testing of procurement platforms for bias, hidden barriers to competition, and unintended discriminatory effects. Automation should not merely reproduce old administrative weaknesses in a more sophisticated form. Transparency is equally important. Suppliers should know the rules that materially affect their participation and should have access to effective review when the system excludes or disadvantages them. So the balance is not achieved by reducing technology. It is achieved by ensuring that every automated step remains explainable, auditable, reviewable, and legally attributable to a responsible public authority.

Dr. Ahmed Mokhtar: Legal judgment should prevail. A smart contract may execute a penalty automatically because a predefined condition has been triggered, but legal liability cannot always be reduced to code. Exceptional circumstances, force majeure, unforeseen events, employer-caused delay, prevention, or other legally relevant facts may alter whether the penalty is actually justified. This is especially important in construction and public contracts, where performance is rarely mechanical. Delays may result from multiple causes, responsibilities may be concurrent, and the contractual position may depend on notices, extensions of time, variation procedures, or the conduct of the employer and engineer. A code-based system may identify that a deadline was missed, but it cannot by itself determine whether the contractor is legally responsible for that delay. My research on smart administrative contracts and digital public procurement has reinforced this concern. Automated execution can improve efficiency, but it should not eliminate legal interpretation or the possibility of human intervention where exceptional facts arise. In practical terms, smart contracts should include a legal “pause” or review mechanism. If the contractor raises a credible exceptional circumstance, automatic enforcement should be suspended long enough for the relevant authority, engineer, dispute board, tribunal, or court to assess the facts and the governing contract. So the code may trigger the process, but it should not have the final word. In my view, code should execute the contract, but law should govern the contract.

Dr. Ahmed Mokhtar: From my experience, the contractual mistake that most often develops into a major dispute is not always a badly drafted clause. Very often, it is the failure to administer a reasonably drafted contract properly. In construction projects, disputes frequently begin with something that appears minor at the time: a variation carried out before formal approval, a delay not notified within the required period, additional work performed without clear valuation, an instruction given informally, or correspondence that does not clearly reserve contractual rights. Months later, when payment, extension of time, liquidated damages, or responsibility for delay becomes disputed, those apparently small procedural failures become central. My professional work in construction contracts, commercial transactions, dispute resolution, litigation strategy, arbitration-related matters, and regulatory compliance has made me particularly conscious of this problem. I would therefore identify poor contract administration and weak contemporaneous documentation as one of the most dangerous recurring mistakes. A strong contractual entitlement can become difficult to enforce if the party cannot prove what happened, when it happened, who instructed it, what its impact was, and whether the contractual notice procedure was followed. A second recurring problem is ambiguity in risk allocation. If responsibility for design changes, site conditions, delays, approvals, or interface risks is not clearly allocated, the dispute may be built into the project from the beginning. My practical advice is simple: parties should treat contract administration as part of legal risk management, not as clerical work. Notices, records, variation procedures, programme updates, payment documentation, and reservation of rights should be managed continuously throughout the project. In construction law, disputes are often won or lost long before anyone enters arbitration or court.

Dr. Ahmed Mokhtar: In a FIDIC-based project, the best protection against avoidable claims begins before the first dispute appears. Employers and contractors should treat contract administration as a core project function from day one, not as something activated only when a problem arises. From my practical work with construction contracts, disputes, and arbitration-related matters, I have found that the strongest protection usually comes from five disciplines: understanding the contract, preserving records, issuing notices on time, controlling variations, and maintaining an accurate programme and payment trail. First, both parties should identify the clauses dealing with notices, extensions of time, variations, payment, claims, and dispute resolution, and then build internal procedures around those requirements. A contractual notice should never depend on memory or informal communication. Second, contemporaneous records are essential. Daily reports, correspondence, programmes, site instructions, approvals, photographs, cost records, and meeting minutes often become decisive when causation and responsibility are later disputed. Third, variations should be documented before execution wherever possible. The instruction, scope, valuation method, time effect, and reservation of rights should be clear. Fourth, programme management must be disciplined. Delay claims are difficult to assess without reliable baseline and updated programmes showing cause, effect, and critical impact. Fifth, payment administration should be transparent and supported by agreed measurement and certification processes. I would also recommend early escalation of disputed issues rather than allowing them to accumulate until project completion. The practical lesson is simple: in FIDIC-based projects, many disputes do not arise because the contract lacks rules, but because the parties fail to use those rules consistently. Good contract administration is therefore not paperwork; it is dispute prevention.

Dr. Ahmed Mokhtar: The choice should not begin with a preference for one dispute mechanism. It should begin with the nature of the dispute, the contract, the urgency, the amount at stake, the need to preserve the project relationship, and the enforceability of the eventual outcome. From my practical work in construction contracts, dispute resolution, litigation strategy, and arbitration-related matters, I have found that no single mechanism is superior in every case. Negotiation should usually be the first step where the commercial relationship is still functioning and the dispute can be resolved without weakening legal rights. It is faster, less expensive, and gives the parties control over the result. Mediation is useful where communication has broken down but both parties still want a commercial settlement with the assistance of a neutral third party. Adjudication is especially valuable in construction projects where a quick interim decision is needed to keep cash flow and performance moving. In FIDIC-based projects, dispute boards can play an important role because they understand the project while it is still ongoing. Arbitration is often appropriate for major technical and cross-border disputes, particularly where confidentiality, specialist decision-makers, procedural flexibility, and international enforceability are important. Litigation may be preferable where urgent court powers are required, where third parties must be joined, where public-law issues arise, or where the contract itself points to the courts. The key point is sequencing. A party should assess the dispute early and decide which mechanism best protects time, cost, evidence, enforceability, and the commercial objective. In serious construction disputes, the best strategy is often not to ask, “Which forum is best?” but rather, “Which route gives the strongest practical outcome for this particular dispute?”

Dr. Ahmed Mokhtar: Law schools should increasingly focus on the abilities that information access cannot replace: legal judgment, critical reasoning, interpretation, ethical responsibility, and the ability to understand facts in context. Students today can retrieve legislation, cases, commentary, and even draft legal arguments within seconds. That is a major advantage, but access to information is not the same as understanding law. The real challenge is knowing which source matters, whether it is authoritative, how competing rules interact, what facts are legally significant, and whether a seemingly persuasive answer is actually correct. My teaching philosophy has always emphasized analytical engagement, problem-based learning, comparative legal thinking, and the ability to apply legal rules to practical situations rather than simply memorising them. I would therefore place much greater emphasis on three areas. First, legal reasoning. Students must learn how to identify the real issue, distinguish strong arguments from weak ones, and justify conclusions rather than merely produce them. Second, professional judgment and ethics. AI can generate options, but it cannot assume professional responsibility for advising a client, balancing competing interests, or deciding when a technically possible course is legally or ethically inappropriate. Third, verification and critical use of technology. My experience reviewing AI-assisted legal and academic outputs has shown how easily confident language can conceal errors in authority, reasoning, citation, or context. So I do not think law schools should compete with AI in providing information. They should teach students how to question information, test it, apply it, and take responsibility for the consequences. The lawyer of tomorrow will not be distinguished by how much law they can retrieve, but by how well they can reason with it.

Dr. Ahmed Mokhtar: The most common weakness I see is not that AI always gives obviously wrong answers. It is that it often gives answers that sound convincing enough to discourage verification. That is dangerous in legal work. From my experience reviewing AI-assisted legal and academic outputs, the recurring problems are usually found in four areas: authority, context, reasoning, and confidence. A system may cite a provision inaccurately, rely on an outdated rule, miss a jurisdictional distinction, or produce a logically smooth answer that does not fit the facts. My work in AI-supported legal review has therefore focused on legal accuracy, regulatory compliance, reasoning quality, citations, terminology, and coherence. The deeper problem is what I would call delegated judgment. Some users begin to treat the AI output as the conclusion rather than as a draft, research aid, or analytical starting point. Once that happens, the lawyer stops asking the questions that matter: Is this authority current? Does it apply in this jurisdiction? Are the facts complete? Is there a conflicting rule? Is the conclusion legally defensible? AI is very useful for structuring issues, comparing positions, identifying research paths, and improving drafting efficiency. But it becomes risky when speed replaces verification. In practical terms, I recommend a simple rule: every AI-generated legal proposition that could materially affect a client, case, contract, or regulatory decision should be independently checked against the original legal source and the actual facts. The real value of AI in law is not that it removes legal judgment. It is that it gives lawyers more time to exercise legal judgment well.

Dr. Ahmed Mokhtar: Both. Future lawyers need to become better users of AI and better critics of it at the same time. Using AI effectively is now becoming part of legal competence. Lawyers who understand how to structure queries, compare outputs, organize large volumes of information, identify research paths, and use AI to improve drafting efficiency will work faster and often more systematically. But technical fluency alone is not enough. My experience reviewing AI-assisted legal and academic work has shown that the more persuasive the output appears, the more important critical judgment becomes. Lawyers must be able to test the source, verify the law, identify missing context, detect overconfidence, distinguish between jurisdictions, and understand when an apparently coherent answer is legally weak. That is why I see the lawyer of the future as having a dual role. First, as an intelligent user of technology. Second, as a professional gatekeeper responsible for ensuring that technology does not lower the standard of legal reasoning. Law schools and professional institutions should therefore teach both practical AI use and AI criticism. Students should learn not only how to generate a legal draft, but how to audit it. They should ask: What source supports this statement? Is it current? Is it applicable here? What fact could change the conclusion? What risk has the system overlooked? So the future lawyer should neither reject AI nor surrender to it. The strongest lawyer will be the one who knows when to use AI, how to question it, and when to disregard its answer entirely.

Dr. Ahmed Mokhtar: If I were designing a course called “Law in the Age of AI,” I would make three subjects essential because they reflect the core challenges already shaping legal education and practice. The first would be AI, public power, and fundamental rights. Students should understand how automated decision-making affects legality, equality, privacy, due process, administrative accountability, and constitutional protection. This connects directly with my research on AI regulation, digital identity, digital rights, predictive government, and automated public decision-making. The second would be data governance, transparency, and responsibility. Future lawyers must understand how data is collected, processed, shared, and used to generate legal or administrative consequences. They should also study questions of explainability, bias, auditability, human oversight, and liability when AI systems cause harm. My work on administrative liability and data governance has reinforced how central these issues are to modern public law. The third would be AI in legal practice and professional judgment. Students should learn how AI can assist legal research, drafting, contract review, compliance, and dispute analysis, but also how to verify outputs, detect errors, preserve confidentiality, and maintain professional responsibility. My experience reviewing AI-assisted legal and academic outputs has shown that this critical dimension is indispensable. I would teach all three through practical cases rather than theory alone. Students should evaluate an automated administrative decision, audit an AI-generated legal opinion, and analyze a contract or dispute involving algorithmic tools. The purpose of such a course should not be to teach students how to “use AI” in isolation. It should teach them how to govern, question, and legally control AI while using it intelligently.

Dr. Ahmed Mokhtar: The difference is profound. A university that merely complies with quality standards usually treats quality as a set of forms, reports, indicators, and periodic accreditation requirements. A university that genuinely builds a culture of quality treats quality as a continuous institutional habit. From my experience in academic quality assurance, accreditation, curriculum development, self-evaluation, benchmarking, and improvement planning, I have seen that real quality begins when standards move from documents into daily decisions. A compliance-based institution asks: Have we completed the required forms? Have we met the minimum indicators? Have we prepared for the external review? A quality-driven institution asks different questions: Are students actually learning what the programme promises? Are assessments measuring the intended outcomes? Are graduates prepared for the labour market? Are faculty using evidence to improve teaching? Are weaknesses being identified early and converted into measurable improvement plans? For me, four elements distinguish a genuine quality culture: leadership commitment, reliable evidence, faculty ownership, and continuous improvement. Quality cannot remain the responsibility of one office. It must be shared across academic departments, committees, administrators, and teaching staff. Another important difference is how the institution reacts to weakness. A compliance-oriented university may try to present the best possible picture before an accreditation visit. A mature quality culture is willing to identify shortcomings honestly, because improvement begins with accurate diagnosis. So I see accreditation as an important tool, but not the ultimate goal. The real goal is to create an institution that evaluates itself continuously, learns from evidence, and improves even when no external reviewer is watching.

Dr. Ahmed Mokhtar: The priority that is most often misunderstood, in my view, is employability. Universities sometimes reduce employability to short-term job placement, while the real objective should be broader: preparing graduates with the knowledge, analytical ability, professional judgment, communication skills, digital competence, and adaptability needed to remain effective as professions change. This is particularly important in legal education. A law graduate may know legislation and doctrine very well, yet still be unprepared for practice if they cannot analyse facts, draft clearly, use technology responsibly, communicate with clients, understand institutional processes, or work across different legal and professional environments. My experience in curriculum development, quality assurance, academic improvement planning, benchmarking, and aligning graduate outcomes with labour-market needs has reinforced this point. Employability should therefore influence curriculum design, assessment methods, clinical learning, internships, digital skills, and continuous feedback from employers and professional bodies. Research, rankings, digital transformation, and internationalisation are all important, but they should support the university’s academic mission rather than become isolated targets. A higher ranking is valuable only if it reflects genuine academic strength. Digital transformation is useful only if it improves learning and institutional performance. Internationalisation matters when it creates meaningful academic exchange and broader competence. So I would not ask whether universities should choose between research, rankings, employability, or internationalisation. The real question is whether these priorities are integrated around student learning and institutional purpose. A university succeeds when its graduates are not only employable on graduation day, but capable of learning, adapting, and remaining professionally relevant years later.

Dr. Ahmed Mokhtar: What separates an outstanding young legal scholar from someone who simply produces many publications is not volume, but intellectual direction, methodological discipline, and the ability to produce work that changes how a legal problem is understood. A strong scholar does more than publish frequently. They build a coherent research agenda. Their work develops around connected questions, deepens over time, and shows a recognizable intellectual identity. In my own research, for example, I have tried to connect public law, digital governance, AI regulation, digital rights, smart government contracting, and administrative responsibility rather than treat each publication as an isolated topic. Quality also depends on method. A serious legal researcher should ask whether the research question is genuinely important, whether the comparative framework is justified, whether primary legal sources have been used accurately, and whether the analysis contributes something beyond description. Another distinguishing feature is independence of thought. Outstanding scholars do not merely repeat prevailing views. They test them, compare them, identify weaknesses, and propose workable alternatives. That is particularly important in fast-changing areas such as AI law, where legal scholarship must often address problems before settled doctrine fully exists. I also believe that a strong scholar connects research with teaching, practice, and institutional reality. My own work in legal education, academic reviewing, quality assurance, and professional legal practice has reinforced the value of research that can move between theory and application. So I would advise young researchers not to chase publication counts. Build a field, develop a voice, protect methodological rigor, and work on questions worth answering. A long publication list may attract attention, but lasting academic value comes from clarity of thought, originality, reliability, and contribution.

Dr. Ahmed Mokhtar: I expect AI to transform most deeply the areas of law that depend heavily on information processing, pattern recognition, documentation, and repetitive decision structures. Legal research, compliance, contract review, due diligence, regulatory monitoring, administrative processing, and parts of dispute analysis will all change significantly. AI can already organize large volumes of material, compare clauses, identify inconsistencies, detect patterns, and support faster legal drafting. My own work on AI-supported legal evaluation and digital governance has shown how quickly these functions are evolving. Public law will also be deeply affected, particularly where governments use AI in licensing, benefits, procurement, inspections, risk assessment, digital identity, and automated decision-making. This is why questions of transparency, accountability, data governance, and human oversight will become even more important. But some parts of law will remain strongly human. Strategic judgment, advocacy, negotiation, ethical responsibility, interpretation of ambiguous facts, assessment of credibility, and balancing competing rights require more than information processing. They require context, experience, responsibility, and sometimes empathy. Judicial decision-making is a good example. AI may assist judges by organizing precedents or identifying relevant issues, but difficult cases often involve proportionality, fairness, constitutional values, and social consequences. These are not merely technical calculations. So I do not expect AI to replace law as a human discipline. I expect it to redistribute legal work. Routine cognitive tasks will increasingly be automated, while the value of human judgment will rise in the areas where law requires responsibility, interpretation, and choice. The future legal professional will therefore be less valuable for retrieving information and more valuable for knowing what to do with it.

Dr. Ahmed Mokhtar: For a young lawyer or researcher seeking an international career at the intersection of law, technology, and public policy, I would recommend building a profile that is broad enough to understand systems, but deep enough to be credible in one or two areas. The first priority is strong legal foundations. Technology changes quickly, but principles such as legality, constitutional rights, administrative accountability, contracts, liability, and due process remain essential. Without that foundation, it is easy to discuss technology in fashionable terms without understanding its legal consequences. The second priority is digital literacy. A lawyer does not need to become a programmer, but should understand how data, algorithms, automated decision-making, AI systems, and digital platforms operate at a practical level. My own work on AI law, digital governance, data governance, and automated public decision-making has reinforced how important this interdisciplinary understanding has become. The third priority is comparative and international thinking. Young researchers should learn to compare legal systems carefully, work with primary sources, and understand how the same technological problem may be regulated differently across jurisdictions. I would also strongly recommend developing research and writing skills, professional English, and the ability to communicate legal ideas clearly to non-lawyers. International careers are built not only on knowledge, but on the ability to explain complex issues across disciplines and cultures. Finally, young professionals should learn to verify information, question technology, and remain intellectually independent. If I had to summarise the advice in one sentence, it would be this: build legal depth, technological literacy, comparative vision, and the discipline to think for yourself.

Dr. Ahmed Mokhtar: One idea I understand very differently today is that law does not shape society through rules alone. It shapes society through institutions, implementation, trust, and the quality of judgment exercised by those who apply it. At the beginning of my career, like many young legal academics, I was naturally drawn to legal texts, principles, and doctrinal structure. Over time, teaching, practice, research, and institutional work showed me that a well-drafted rule can still produce weak outcomes if institutions are ineffective, procedures are unclear, enforcement is inconsistent, or people do not trust the system. My experience in legal education, quality assurance, public law, corporate legal practice, dispute resolution, and emerging areas such as AI governance has reinforced this view. I have seen that the same legal principle can function very differently depending on administrative culture, institutional capacity, professional ethics, and the availability of meaningful review. Technology has made this lesson even clearer. An advanced digital system may improve efficiency, but if accountability, transparency, and human judgment are weak, technology can simply automate institutional problems rather than solve them. So today I see law less as a collection of rules and more as an ecosystem. Text matters, but so do institutions, procedures, professional competence, and public confidence. If I had to express the lesson in one sentence, it would be this: the true strength of law is not measured by how sophisticated its rules are, but by whether people can rely on those rules to produce fair, predictable, and accountable outcomes in real life.