ABSTRACT

1 INTRODUCTION

Artificial intelligence (AI) is a real and effective technology whose applications, in some areas, exceed human capabilities. It is capable of reasoning, perception, problem-solving, and even autonomous learning. AI systems can integrate and utilise various advanced tools and devices to perform complex tasks efficiently.

However, the development of self-learning AI systems remains a major challenge. Many questions arise regarding their training, ethical use, and responsibility. At present, few mechanisms exist to adapt AI to specific cultural or linguistic contexts, particularly in environments where AI systems must recognise and respond to unique local characteristics.

To ensure the safe and ethical use of AI, it is essential to develop robust systems that promote transparency, accountability, and fairness. Such systems would facilitate automatic control while minimising potential risks.

This issue is particularly important in light of recent technological advancements, as AI increasingly influences contractual frameworks and civil liabilities, especially in regions where Arabic is the dominant language, such as Egypt and the United Arab Emirates. There is an urgent need for individuals to understand the basic concepts of AI and related technologies, especially in this rapidly evolving digital environment.

A review of Egyptian and Emirati legislation reveals significant shortcomings in addressing the legal challenges posed by AI-based negotiation mechanisms. Neither legal system includes explicit provisions recognising the legal status of AI systems or robots, nor do they provide clear definitions of the rights and obligations applicable to such artificial entities. Moreover, both jurisdictions lack a dedicated regulatory or supervisory body to oversee the operation of AI agents or to ensure accountability for civil liability arising from their role

in contractual negotiations. This stands in contrast to the more advanced European approach, which has moved toward establishing independent oversight bodies and imposing greater transparency obligations. In this regard, it is worth noting that the European Law Institute has issued a set of guiding principles on automated decisionmaking within the European Union.¹

Within the European legal framework, it is essential to distinguish among three principal instruments. The EU Artificial Intelligence Act (AI Act), enacted as Regulation (EU) 2024/1689,² establishes a comprehensive risk-based legal regime for AI systems. It classifies such systems into four levels of risk—unacceptable, high, limited, and minimal—and imposes corresponding obligations on providers and deployers, including conformity assessments and transparency requirements.

The EU Data Act (Regulation (EU) 2023/2854)³ focuses on enabling fair access to and use of data within the European data economy.⁴ It forms a critical part of the legal foundation for smart contracts, especially in data-sharing agreements. These smart contracts are subject to essential legal requirements, such as auditability, access control, and conformity certification.

The AI Liability Directive⁵ (AILD)—originally proposed to harmonise non-contractual civil liability rules for AI across the Union—seeks to simplify the claimant’s burden of proof by facilitating access to evidence and introducing rebuttable presumptions of causality.⁶ However, its legislative future remains uncertain.

The main research question guiding the study is: How can Arab regulatory frameworks bridge the gap in civil liability arising from AI-driven negotiations by leveraging the European model, while accounting for the region's legal specificities?

This study seeks to address the pressing legal challenges arising from the use of AI in contractual negotiations by formulating solutions consistent with prevailing legal traditions and legislative frameworks. Its objectives are fourfold: first, to emphasise the significance of the pre-contractual stage and underscore the legislator’s duty to regulate it, particularly in determining the scope of civil liability associated with this phase; second, to examine the legal implications of AI in negotiation and the challenges it generates; third, to propose a balanced legal framework suitable for the responsible deployment of AI in contractual bargaining; and fourth, to explore the attribution of liability for harms caused by AI. 2 METHODOLOGY OF THE STUDY

2 METHODOLOGY OF THE STUDY

This study adopts a comparative analytical approach, aiming to examine the legal framework governing civil liability for AI-related errors during the negotiation phase. The analysis focuses on the fundamental legal theories of tort and contractual liability as established in Arab civil laws, particularly those of Egypt and the UAE. This comparative approach is also employed to examine this situation in relation to recent legislative and judicial developments in the European Union, specifically the AI Liability Directive and the Data Act, which address these issues more directly than the AI Act. 3 AI CONCEPT AND RULE IN CONTRACTUAL NEGOTIATION

3 AI CONCEPT AND RULE IN CONTRACTUAL NEGOTIATION

AI is a branch of computer science focused on developing systems that simulate human behaviour and decision-making with varying degrees of autonomy.⁷ It operates through software or integrated devices such as robots and self-driving cars. AI processes structured, semi-structured, and unstructured data to analyse environments and solve complex problems. Its applications include digital assistants, facial recognition, and autonomous machines like drones. Due to their efficiency and adaptability, these technologies are increasingly utilised in medicine, economics, and defence.⁸

Contractual negotiation involves the exchange of proposals, opinions, studies, and legal consultations between parties aiming to reach a mutually beneficial agreement. It plays a crucial role in determining the terms and conditions of contracts, ensuring that the interests of all parties are balanced and that business transactions are successful. Negotiations

require individuals with specialised skills and knowledge and take place across various fields, including hospitals, offices, and legal consultancies.⁹

AI can enhance contractual negotiations through data analysis, enabling parties to assess market conditions, understand the needs of each party, and identify strengths and weaknesses. It can also evaluate risks, determine optimal negotiation paths, and propose creative solutions. By automating repetitive tasks, AI saves negotiators time, enabling them to focus on more important aspects of the process. However, some scholars contend that AI will not fully replace humans in negotiations, as human input is essential for understanding the other party’s position.¹⁰

Several companies, including IBM, Salesforce, and Alibaba, utilise AI technologies (e.g., Watson, Einstein, and Alibaba DAMO) to assist in negotiations. While AI is increasingly seen as an inevitable force in legal and judicial matters, scholars emphasise that it presents challenges that must be addressed legislatively and technically. They caution that AI should remain under human control to mitigate risks and ensure its ethical integration into society.¹¹ 4 LEGAL CHALLENGES RELATED TO CONTRACTUAL NEGOTIATION THROUGH AI TECHNOLOGIES

4 LEGAL CHALLENGES RELATED TO CONTRACTUAL NEGOTIATION THROUGH AI TECHNOLOGIES

The widespread adoption of artificial intelligence has generated significant legal challenges, particularly with respect to its inherent risks, as well as broader concerns in the domains of research and innovation. This tension largely stems from the accelerating pace of technological advancement, which often outstrips the ability of legal frameworks to adapt, thereby exacerbating these challenges. In this context, the present study focuses on civil liability for damages arising from errors committed by artificial intelligence systems during the contractual negotiation process.

4.1. Civil Liability for Damages Caused by Artificial Intelligence

The issue of civil liability arising from AI errors in negotiations is a recent and complex matter that requires careful consideration from both legal and technical perspectives.

In general, civil liability is defined as the obligation under which a person is liable for remedying the damage incurred by another person due to the acts performed by the

former, their subordinates, or things for which the former is liable.¹² It can also be defined as “the person’s obligation to compensate the damage he caused to another person because of violating an obligation represented in infringing the victim or third parties in whatsoever manner.” ¹³

In the context of AI-related damages, civil liability refers to the liability of AI for compensating the damage incurred by the victim or a third party as a result of the operation or decision-making of an AI system.¹⁴

With the growing use of AI in fields like healthcare, civil liability for AI-related errors has become increasingly important—for example, when a robot causes harm to a patient. In such cases, the responsible party must compensate for financial and moral damages, regardless of fault. This strict liability principle applies to institutions using AI, such as hospitals, and serves as a warning to manufacturers of potentially dangerous technologies.

Since the mid-2010s, the European Parliament has shown heightened interest in civil liability arising from AI applications, particularly those involving robots and autonomous vehicles. On 16 February 2017, the Parliament adopted a resolution on Civil Law Rules on Robotics, calling for the development of new legal frameworks that account for the difficulty of proving software errors and the challenges arising from system autonomy in decision-making.¹⁵ This was followed by another resolution on 20 October 2020 (2020/2014(INL)), which recommended the establishment of a comprehensive civil liability system and a balanced compensation scheme capable of addressing damages resulting from the use of AI technologies, recognising the inadequacy of traditional rules based on proving fault and causality.¹⁶

Member States have responded to these recommendations in various ways. In 2021, Germany adopted a specific law on autonomous driving, requiring the presence of a "technical supervisor" (technische Aufsicht) in vehicles and mandated the installation of a "black box"-like device to record driving data for determining liability in the event of an accident. This law represents a practical national model for addressing the evidentiary

challenges in AI-related incidents, complementing European discussions on the need for a unified framework.¹⁷

In a broader context, the European Union adopted the Artificial Intelligence Act (AI Act) in July 2024, marking the first binding horizontal European legislation regulating AI technologies. This law is based on a risk-classification methodology, prohibiting systems with "unacceptable risk" and subjecting high-risk systems to stringent technical and procedural obligations, including risk management, mandatory human oversight, maintaining transparent operational records, and reporting malfunctions. The Act also provided for the establishment of national regulatory authorities to supervise compliance, while imposing limited transparency requirements for low-risk systems.¹⁸

Alongside these developments, the EU Data Act, enacted in 2023, represent a parallel regulatory step aimed at ensuring access to and defining usage rights for data generated by connected devices, thereby providing a supportive legal environment for AI development by facilitating industrial data sharing.

Regarding compensation and liability, the European Commission proposed the AI Liability Directive in 2022, aiming to alleviate the burden of proof for victims by introducing mechanisms such as legal presumptions of causality. However, this proposal faced broad political and legislative disagreements, leading to its withdrawal from the Commission's work programme in 2025, leaving a partial legislative gap in compensation for AI-related damages.¹⁹

These European developments reflect the European legislator's awareness of the complex nature of AI and the need for innovative legal tools that strike a balance between the requirements of technological innovation and the protection of fundamental rights and contractual interests. Despite such progress, Arab experiences remain more focused on formulating national strategies and general policy frameworks rather than building detailed legislative systems.

In Egypt, the government launched the National Strategy for Artificial Intelligence and enacted the Personal Data Protection Law in 2020; however, issues of civil liability for AI damages remain subject to traditional rules in the Civil Code and Consumer Protection Law.²⁰ In the

UAE, the UAE Strategy for Artificial Intelligence 2031 was launched, and Federal Law No. (45) of 2021 on the Protection of Personal Data was issued, while legislative policy relies on regulatory sandboxes to test systems before widespread deployment, without a specific law yet regulating civil liability for AI-related damages.²¹

An analytical review of the UAE and Egyptian legislative frameworks reveals several fundamental gaps that hinder their ability to address the legal challenges arising from contractual negotiations conducted through AI systems. Chief among these are the absence of explicit legal provisions recognising the legal status of AI systems or robots, and the lack of a precise statutory definition delineating their rights and obligations within contractual frameworks.²²

Moreover, both jurisdictions lack a dedicated regulatory framework or specialised oversight body responsible for supervising the performance of AI agents or ensuring their compliance with civil liability rules during negotiations. This stands in contrast to the European experience, which has pioneered practical models for establishing independent supervisory entities and linking the use of AI to clear legal duties concerning transparency and accountability.

In light of these findings, the study recommends that UAE and Egyptian lawmakers undertake the following steps:

1. Adopt clear and specific legal definitions of intelligent systems and delineate their contractual use within the Civil Transactions Law.
2. Develop dedicated legal mechanisms to impose liability on AI developers and users for harm resulting from automated actions.
3. Establish specialised regulatory bodies—or dedicated units within existing authorities—to oversee smart contracts and assess the legal performance of the algorithms employed.

The study also recommends the gradual implementation of these mechanisms, drawing on recent European models, to ensure a safe transition toward an AI-driven contractual environment while minimising the regulatory gap.

4.2. Civil Types of Damages Resulting from AI Errors in Contract Negotiations

This review is supported by Chopard and Musy, who argue that AI systems are being increasingly used to aid in the diagnosis and treatment of diseases, thereby reducing the risk of medical errors. They note that such systems also influence the determination and allocation of compensation among doctors and producers of AI systems in cases where patients suffer harm.²³ Conversely, other scholars contend that determining liability in AI-related defects could be complex due to the involvement of multiple stakeholders. Responsibility becomes difficult to establish when defects arise from different sources, such as defective training data, algorithmic biases, or inappropriate system design.²⁴

When AI is employed in contract negotiation or dispute resolution, its errors can result in significant financial or legal harm—such as costly decisions, broken agreements, project delays, or unexpected expenses. These mistakes may lead to contractual breaches or tort liability arising from the AI’s actions.²⁵

1. Operational Damages: Errors by AI systems during negotiations could result in delays of projects and business transactions, leading to additional costs and lost opportunities. This results in weakening the company's competitive position and negatively affecting its reputation.²⁶
2. Legal Damages: In addition to operational damages, AI errors in negotiation can result in serious legal problems, including contractual violations or torts (harmful acts), which may lead to expensive judicial disputes.
3. Financial Damages: When AI errors occur in negotiating contracts, this can result in substantial financial losses to the negotiating parties, which could include concluding unprofitable or unfavourable agreements, missing opportunities for profit and business growth.

Civil liability for AI errors in negotiations is established on a group of main legal pillars, foremost among them being the principle of tort (harmful act), which requires compensation for damage resulting from a harmful act or negligence by the entity responsible for the AI.²⁷

This liability is grounded in the principles of contractual liability, under which the entity in charge of AI is obliged to exert due diligence to prevent such damages. In the event of violating such an obligation, it shall compensate the victims for the losses they incur or the damages resulting from the AI errors during negotiations.²⁸

4.3. Conditions for Civil Liability for Errors Committed by Artificial Intelligence during Negotiations

by Artificial Intelligence during Negotiations

Civil liability for errors committed by AI during negotiations arises upon the fulfilment of the following condition:²⁹

1. Incurrence of Actual Damage: Civil liability cannot be established unless the AI error results in actual damage to the negotiating parties, whether financial or moral. The damage must be direct and causally linked to the AI’s error.³⁰
2. Commission of an AI Error: Civil liability arises when the AI system commits an error while performing its negotiation functions. The error must stem from deficiencies in programming or AI performance, rather than from events constituting force majeure or other factors beyond the control of the entity in charge.³¹
3. Causation Relationship: There must be a causation relationship between the AI error and the damage incurred by the negotiating parties. In other words, the damage must be the direct and inevitable outcome of the AI error in negotiation and may not be the result of any other factors.³²

In defining the party civilly liable for compensating damages resulting from the use of AI in contractual negotiations, liability may fall upon one or more of the following:³³

1. Liability of the Robot Manufacturer and Programmer: The manufacturer, developer, or programmer of an AI system may incur civil liability where damage arises from defects in design, errors in programming, or negligence in the manufacturing or development process. In such cases, liability is generally governed by product liability laws, which impose a legal obligation on manufacturers to ensure that their products are safe, effective, and free from defects.³⁴
2. Liability for Use (End-User Responsibility): Where harm results from improper use of the AI system—such as failure to adhere to usage guidelines or intentional misuse—the end user may be held liable. In such instances, liability is determined in accordance with the general principles governing fault-based liability for misuse or negligence.³⁵
3. Contracts and Agreements: Contracts governing the use of AI usually include provisions determining the liability for the damages resulting from the AI errors. Such provisions may release the manufacturer from liability in some cases or limit the amount of compensation. Therefore, such contracts must be carefully reviewed to determine the liable party in case of error that results in damage to third parties or to determine the manner of dividing the liability among the different parties.³⁶
4. Training and Maintenance: Liability may be borne by the user if the error is the result of the lack of training or lack of periodical maintenance of the AI.
5. Determining the Error: In case of error, it is necessary to determine whether the cause lies in programming, a technical error, or user misuse. This requires a technical investigation, involving experts in technology and software.
6. Insurance: With the rapid progress in the field of AI and robotics, legislation can develop to include specific provisions related to the liability for AI errors. For example, the expected laws can consist of existing or providing special insurance to cover the damages resulting from the use of AI, which could add a protection layer for the users. Moreover, AI has a significant impact on the insurance industry and poses a future challenge in light of the potential errors that could occur.³⁷

In general, determining civil liability in this context is dependent on identifying the primary cause of the error and the manner in which local laws deal with such lawsuits. Given the complexity of these issues, it is also necessary to consult specialised legal experts to obtain accurate and context-specific advice regarding liability for AI-related damages.

4.4. Victim’s Obligations in Civil Claims Arising from AI Errors in Negotiation

In civil litigation concerning AI-related errors during contractual negotiations, the claimant is subject to specific procedural and evidentiary obligations.

First, the victim must comply with strict procedural time limits for initiating legal action, as failure to observe statutory deadlines often results in the forfeiture of the right to compensation.³⁸

Second, the burden of proof rests on the claimant, who must demonstrate both the existence of an AI malfunction and establish a direct causal nexus between the system’s error and the harm incurred. Given the autonomous and opaque nature of AI decisionmaking, this requirement represents a significant legal challenge.

Third, the claimant is required to provide sufficient evidence of the alleged harm, which may include financial statements, contractual records, and other forms of documentary evidence. Furthermore, technical documentation and expert testimony are often essential to establish whether the damage was attributable to a programming deficiency, a system malfunction, or improper human use.

Recent scholarship has increasingly emphasised the necessity of procedural innovations, such as evidentiary presumptions and reversed burdens of proof, to effectively balance victims’ rights with the complexities of AI accountability frameworks in the European context.³⁹

The methods by which parties may deny civil liability for AI-related errors in negotiation have given rise to several legal questions, including whether the AI user can rely on the foreign cause.⁴⁰

The concept of foreign cause encompasses the urgent or sudden accidents, force majeure events, acts of third parties, acts of the victim, or technical incidents such as breakdowns and viruses affecting AI systems. Under certain conditions, the AI user may rely on such arguments to deny liability. This possibility finds legal support in Article 373 of the Egyptian Civil Code⁴¹ and Article 287 of the UAE Federal Civil Code.⁴²

The European Parliament (EP) has paid particular attention to the issue of civil liability for damages caused by AI systems, including AI software embedded in robots and autonomous driving cars. On 17 February 2017, the EP adopted a series of recommendations related to the civil liability for the damages incurred by third parties.⁴³

These recommendations highlighted two major challenges:

1. The difficulty of attributing error to AI systems under traditional civil liability frameworks, which typically require the establishment of human fault or negligence as a precondition for civil liability.

2. The limitations of holding AI software liable for cases in which AI can make independent and subjective decisions. In such situations, it becomes problematic to identify a “defect” that caused such damage and the causal link between the assumed defect and the resulting damage.⁴⁴

Accordingly, the European Parliament concluded that the general rules of civil liability are insufficient for addressing the damages caused by AI software and applications. It also urged enacting a special legal framework to accommodate and regulate them in proportion to the nature of AI applications.⁴⁵

Among the legislative responses to this call, Germany’s 2017 amendment to its Road Traffic Act stands out. This legislation introduced specific rules governing the civil liability of autonomous cars, including the following provisions:

1. The driver must be present in the vehicle at all times while it is in motion.

2. The driver must retain control of the vehicle when the AI system prompts manual intervention, particularly when the system requires that the driver take over the steering wheel.

3. Every autonomous vehicle must be equipped with a black box, similar to those in aeroplanes, to record specific data, including the vehicle’s itinerary and the driver’s control status at the time of an incident—specifically, whether the accident occurred while the vehicle was under manual control or autonomous operation. If the accident occurs while the vehicle is operating during autonomous mode, the liability shall be borne by the car manufacturer. However, if the accident occurs due to the driver’s failure, for instance, to take control despite receiving notifications and warnings from the AI system, the driver shall bear the liability.⁴⁶

4.5. Critical Reflections on Regulatory Complexity and Comparative Prospects

Amid the rapid evolution of AI and its growing role in automated contractual negotiations, the European Law Institute (ELI) issued its 2022 Principles on AI, emphasising the protection of the right to human review of automated decisions and the necessity of preventing a denial of access to justice arising from reliance on intelligent negotiation systems.⁴⁷ These guidelines further advocate for the modernisation of traditional legal categories, particularly the notion of “product” to include intelligent software, thereby aligning with evolving approaches to product liability in light of increasing automation.⁴⁸

While these principles provide a valuable theoretical foundation for AI governance within European private law, their scope remains largely Eurocentric and insufficiently tailored to the specificities of non-Western jurisdictions. This underscores the need for comparative research, particularly within Arab legal systems, to assess the adaptability of these principles in light of domestic legal traditions, regulatory frameworks, and socio-cultural constraints. Recent scholarship stresses that legal responses to AI must avoid a mere transplantation of European models, and instead develop context-sensitive frameworks capable of addressing local needs while engaging with global standards of AI governance.⁴⁹

This research trajectory represents a crucial first step toward establishing an Arab perspective on AI-related liability in contractual negotiations, laying the groundwork for a comparative legal framework that balances technological innovation with the protection of fundamental rights.⁵⁰

4.6. Distinguishing Contractual and Tortious Liability in AI-Related Cases

In AI-related disputes—particularly within contractual negotiation contexts—the distinction between contractual and tortious liability is of fundamental importance due to the technical and operational complexity of intelligent systems. Contractual liability arises where a contractual relationship exists between the user or injured party and the developer or operator, entailing obligations such as performance or product safety. In contrast, tortious liability applies in the absence of such a relationship and may be based on fault, negligence, or, in some cases, strict liability—especially with high-risk AI systems.

The European Law Institute emphasises that traditional legal frameworks are no longer sufficient and calls for extending tortious liability to cover harm caused by high-risk AI, even in the absence of a contractual relationship.⁵¹ Similarly, Cogen et al. contend that tortious liability in this context necessitates a reconsideration of fault and evidentiary standards—potentially shifting the burden of proof or introducing legal presumptions to facilitate claims.⁵²

Recent studies⁵³ highlight the practical overlap between contractual and tortious liability, particularly in smart or long-term contracts involving both human and automated elements. This overlap necessitates a redefinition of the conceptual and legislative boundary between the two regimes.

Effective legal regulation of civil liability for AI errors cannot rely solely on classical doctrines. Instead, a hybrid legal framework is needed—one that accommodates the unique characteristics of intelligent systems while ensuring meaningful protection for both contracting parties and third parties.

4.7. Scope of Damage Covered Under Product Liability

In the context of applying product liability rules within the European Union to AI technologies—particularly in contractual negotiations—the delineation of compensable harm emerges as a pivotal issue for ensuring a balance between adequate protection of victims and avoiding disproportionate legal burdens on producers. The recent reform introduced by Directive (EU) 2024/2853 on liability for defective products represents a fundamental update, specifying in an exhaustive manner the categories of compensable damage: death or personal injury, including medically recognised psychological harm; damage to property owned by natural persons (excluding the defective product itself and property used exclusively for professional purposes); and destruction or corruption of data, provided that the data is not used for professional purposes.⁵⁴

The Directive further clarifies that “pure economic loss,” as well as harms linked to privacy violations or discrimination, do not in themselves give rise to liability under this framework, although such harms may be addressed under other liability regimes at the national level. This relatively narrow definition reflects the European approach of facilitating effective redress for individuals affected by defective digital products and software—including AI systems—without transforming product liability into a catch-all mechanism for compensating every form of immaterial or purely economic loss.⁵⁵

From a comparative perspective, recent legal scholarship underscores that the reform of product liability rules was driven by the increasing complexity of digital products, supply chains, and the integration of software and machine learning components, while maintaining the logic of strict liability for producers. The scope of compensable damage was deliberately circumscribed to preserve legal certainty and prevent “liability inflation” that could deter innovation. At the same time, evidentiary burdens have been relaxed in favour of claimants, introducing presumptions of defect and causation to mitigate the technical difficulties of proving harm in cases involving AI technologies.

By contrast, Arab civil law systems adopt a broader approach. In Egyptian law, tort liability is founded on the elements of fault, harm, and causation,⁵⁶ with wide recognition of compensation for both material and moral damages, without the strict categorical limitations found in the EU framework. Similarly, the UAE Civil Transactions Law⁵⁷ explicitly provides for compensation of both material and moral

harm, with Article 293 expressly recognising moral damages and extending compensation to the victim’s heirs in specific cases.⁵⁸

This structure results in a broader remedial scope in Egypt and the UAE than under the European product liability regime, encompassing moral harm and, in practice, certain forms of economic loss, albeit subject to judicial interpretation and doctrinal limitations. Consequently, the transposition of EU product liability rules into Arab jurisdictions requires caution, as the substantive scope of compensable harm and the underlying policy objectives differ significantly: while the European framework is narrowly tailored to protect natural persons under a specialised product liability regime, Arab civil law systems operate within general liability frameworks that are more expansive in their remedial reach.⁵⁹

5 CALCULATING THE COMPENSATION FOR THE DAMAGES RESULTING FROM THE AI ERRORS IN NEGOTIATION

When a defendant fails to rebut allegations of liability for harm caused by an AI system during contractual negotiations, courts will ordinarily order the defendant to pay compensation commensurate with the loss sustained by the claimant. Assessing damages in such cases requires a careful and multifaceted exercise.

First, courts must identify and quantify direct and indirect economic losses, including lost profits (lucrum cessans), additional costs reasonably incurred by the injured party to mitigate or remedy the harm, and losses arising from frustrated or rescinded contracts that resulted from the AI malfunction.

Second, non-pecuniary harms—commonly described as moral damages—must be examined where relevant; such harms may encompass reputational injury, loss of goodwill, and the adverse commercial consequences arising from client attrition or the collapse of strategic relationships attributable to the AI failure.

Third, the evidentiary process necessarily demands both conventional documentary proof (such as financial records, contracts, correspondence) and technical proof, including system logs, incident reports, forensic analyses, and expert testimony on software behaviour and fault.⁶⁰

Insurance plays a dual role in this ecosystem. On one hand, civil-liability insurance— whether as bespoke AI performance policies or as extensions of existing technology-E&O/cyber covers—can provide direct compensatory relief to victims and spread residual risk across underwriters,⁶¹ thereby reducing the immediate financial exposure of developers and deployers. On the other hand, well-designed insurance markets can foster responsible innovation by incentivising appropriate governance, testing, and maintenance practices; insurers may require conformity with best practices as underwriting conditions.⁶²

Nevertheless, insurance solutions have limitations: apportioning liability among developers, vendors, and end-users is often technically and contractually complex; the scarcity of historical loss data for novel AI failure modes complicates underwriting and pricing; and rapid technological change risks producing coverage gaps unless policy wordings and regulatory guidance evolve in tandem with technology.⁶³

From a legal-policy standpoint, improving victim protection in AI negotiation scenarios requires a three-pronged approach:

1. Clearer substantive liability rules, including calibrated rules on causation and presumptions where appropriate.
2. Robust evidentiary and technical infrastructures for incident analysis and attribution.
3. Development of insurance mechanisms and regulatory incentives that both compensate victims and promote risk-reducing behaviour by market participants.⁶⁴

Under Egyptian law, tort liability rests on the traditional triad of fault, damage, and causal link,⁶⁵ with courts routinely recognising both material and moral damages (including loss of reputation and consequential commercial losses), provided they are proven, foreseeable, and proximate to the wrongful act. Consequently, Egyptian practice generally permits recovery for lost profits and reputational harm when such losses can be substantiated and causally linked to the defendant’s conduct or the malfunctioning system.⁶⁶

The UAE legal framework similarly follows a general tort model,⁶⁷ recognising compensation for both material and moral injury (see Article 293 regarding moral

damages). Thus, both jurisdictions operate within broad remedial systems that—at least doctrinally—allow compensation for economic and non-economic harms arising from AI failures, subject to the usual constraints of proof, foreseeability, and causation. These features contrast with the evolving European product-liability approach, which narrowly defines compensable damage under the product liability instrument while relying on other regimes for purely economic or privacy-related harms. Accordingly, transplanting European product-liability rules into Egyptian or Emirati legal contexts would require careful calibration to account for the more expansive remedial traditions and evidentiary practices in those jurisdictions.⁶⁸

6 FINDINGS

6.1. Results

The analysis reveals that the integration of AI into contractual negotiations presents significant legal challenges, primarily due to the absence of explicit legislative regulation in Arab jurisdictions and the lack of legal recognition for the autonomous features of AI systems. Egyptian and Emirati civil codes continue to rely on traditional liability structures—built on fault, harm, and causation—without providing tailored provisions for AI-driven decision-making. This results in regulatory uncertainty concerning the allocation of liability among developers, operators, and end-users.

A comparative examination of the European Union reveals a more advanced, though still evolving, framework. Instruments such as the AI Act (Regulation (EU) 2024/1689), the Data Act (Regulation (EU) 2023/2854), and the ongoing debate around the AI Liability Directive collectively seek to mitigate evidentiary burdens on victims, impose obligations of transparency and human oversight, and narrowly define compensable damages under product liability rules. The EU approach strikes a balance between protecting victims and safeguarding innovation, introducing rebuttable presumptions of defect and causation to address the opacity of AI systems.

By contrast, Egyptian and Emirati legal systems adopt broader remedial traditions. Both jurisdictions allow compensation for material and moral damages, including reputational harm and lost profits, without the restrictive categories found in the EU framework. However, the absence of statutory definitions of “intelligent systems,” the lack of dedicated supervisory bodies, and the reliance on general civil code provisions hinder their ability to address the unique challenges posed by AI errors in negotiation.

Consequently, the comparative analysis underscores three key findings:

1. Regulatory Gap: Arab jurisdictions lack specialised legislation to address AI liability, in contrast to the EU’s incremental regulatory reforms.
2. Doctrinal Tension: The strict liability logic of European product law diverges from the broader remedial approach in Arab civil law, complicating any direct transplantation of rules.
3. Evidentiary Complexity: Across both systems, proving causation and fault in AIrelated harm remains a fundamental obstacle, necessitating novel legal and technical mechanisms.

6.2. Recommendations

Based on these findings, the study proposes several normative measures:

1. Adopt precise legal definitions of AI and intelligent systems within civil codes, ensuring clarity in determining rights, duties, and liability.
2. Establish specialised regulatory or supervisory authorities in Arab jurisdictions to oversee the use of AI in contractual contexts, drawing inspiration from the EU model of independent oversight bodies.
3. Develop hybrid liability frameworks that integrate contractual and tortious doctrines with calibrated presumptions of defect and causation, thereby easing the burden of proof for victims while maintaining fairness to developers and users.
4. Strengthen insurance mechanisms tailored to AI-related risks, both as compensatory instruments and as tools for incentivising responsible AI governance and risk management practices.
5. Ensure context-sensitive legal transposition by avoiding wholesale adoption of European models and instead designing frameworks that respect Arab legal traditions while engaging with global standards of AI regulation.

Together, these recommendations aim to bridge the current regulatory gap, enhance legal certainty, and promote a balanced framework that simultaneously safeguards victims, supports innovation, and ensures accountability in AI-driven contractual negotiations.

FOOTNOTES

REFERENCES

AUTHORS INFORMATION

RIGHTS AND PERMISSIONS

EDITORS

ABOUT THIS ARTICLE

DETAILS FOR PUBLICATION

AI DISCLOSURE STATEMENT

АНОТАЦІЯ