{"id":41498,"date":"2026-07-03T14:55:29","date_gmt":"2026-07-03T11:55:29","guid":{"rendered":"https:\/\/afteegypt.org\/?p=41498"},"modified":"2026-07-03T15:10:48","modified_gmt":"2026-07-03T12:10:48","slug":"41498","status":"publish","type":"post","link":"https:\/\/afteegypt.org\/en\/research-en\/2026\/07\/03\/41498-afteegypt.html","title":{"rendered":"An Assessment of Artificial Intelligence Strategies in Egypt: Between Challenges and Aspirations"},"content":{"rendered":"

Contents<\/span> <\/strong><\/h4>\n

Methodology<\/strong><\/p>\n

Introduction <\/strong><\/p>\n

Section One: AI Governance: Technical Specifics and Challenges of Prioritisation and Regulation Worldwide<\/strong><\/p>\n

Section Two: The National Artificial Intelligence Strategy: Between Reality and Aspirations<\/strong><\/p>\n

Conclusion and Recommendations <\/strong><\/p>\n

 <\/p>\n

Methodology<\/span> <\/strong><\/h4>\n

In its analysis and assessment, the paper drew on a desk review of a number of international reference frameworks on the governance of artificial intelligence, in particular the international outputs issued by the Organisation for Economic Co-operation and Development (OECD) as the lead coordinator and supporter of governments\u2019 efforts worldwide to adopt artificial intelligence technologies. These frameworks included the OECD\u2019s outputs on the application of artificial intelligence (AI) within the scope of government work \u2013 which is the main objective of Egypt\u2019s National AI Strategy \u2013 and which, in turn, were based on a survey of experiences in applying artificial intelligence in government services worldwide, and an evaluation of these applications.<\/p>\n

The paper also based its assessment on the OECD\u2019s recommendations on responsible and trustworthy artificial intelligence (Principles of Trustworthy AI). Furthermore, the paper drew upon another reference framework: the European Union\u2019s Artificial Intelligence Act (EU Regulation 1689\/2024), This choice was made because it is, to date, the most comprehensive and integrated regulatory framework globally, having been enacted as legislation designed to promote the adoption of artificial intelligence technologies whilst protecting citizens\u2019 personal data and upholding human rights and fundamental freedoms.<\/p>\n

The paper also drew on a review of public policy studies in the field of artificial intelligence regulation worldwide, examining the challenges they raise and the lessons learnt from various governance experiences around the world. Furthermore, the paper surveyed and analysed relevant Egyptian national frameworks relating to artificial intelligence, such as the National AI Strategy (in its first and second editions), the National Guidelines for Generative AI, and the Egyptian Code of Ethics for Responsible Artificial Intelligence. The analysis and evaluation also drew on other frameworks that complement the AI agenda within the telecommunications and information technology regulatory sector, such as the Digital Egypt Strategy, the Cloud Computing Strategy, the Open Source Software Strategy, and the Open Data Strategy, as well as a review of the outputs and efforts of the Egyptian Centre for Artificial Intelligence in implementing and reviewing the National AI Strategy.<\/p>\n

\u00a0<\/strong><\/p>\n

\u00a0<\/strong><\/p>\n

Introduction<\/strong><\/span><\/h4>\n

AI is attracting growing attention both globally and in Egypt, given the rapid development of this technology, the increasing reliance on it across a wide range of sectors, and the huge returns this sector is expected to generate. The Egyptian government began to take an institutional interest in AI by launching the National Council for AI in 2019, followed by the publication of the first version of the AI Strategy in 2021, with a second version due in 2025. This was followed by attempts to draw up proposed guidelines for dealing with AI in Egypt, the most recent of which were the National Guidelines for Generative AI in March 2026. In parallel, a number of draft bills have been proposed by members of parliament to regulate and govern AI in Egypt[1]<\/sup><\/a> , whilst the government is also working on drafting a similar bill.[2]<\/sup><\/a><\/p>\n

This paper aims to contribute to the ongoing discussions on public policy for the regulation of the AI sector, and the associated protection of digital rights, the right to privacy and the guarantee of the free flow of information. To this end, the paper examines the National AI Strategy \u2013 in both its versions \u2013 and related guiding frameworks such as the Guidelines for the Regulation of Generative AI and the Egyptian Code of Ethics for Responsible AI. This requires an understanding of the legal and regulatory treatment of AI, within the context of Egypt\u2019s communications and information infrastructure, the nature of the proposed technologies, and the target sectors.<\/p>\n

The paper based its assessment on the findings and recommendations of the OECD regarding the governance of AI, particularly in government work. The most significant and recent of these was a publication titled \u201cGoverning with AI\u201d in 2025, which drew on a survey of 200 AI adoption cases across 12 government sectors in various countries.[3]<\/sup><\/a> This publication is notable for its focus on monitoring and analysing the efforts of governments worldwide; in producing it, and in highlighting the lessons learnt from these implementation experiences, the OECD drew on information provided by the governments themselves. This edition forms part of a wider project sponsored by the organisation, entitled \u2018Governing AI\u2019, which aims to support governments in adopting AI in a responsible and secure manner. The report also drew on policy papers prepared by independent researchers for the OECD\u2019s AI Policy Observatory, which address the challenges of adopting AI in the delivery of public services. The paper also addresses the OECD\u2019s recommendations, known as the \u2018Principles of Trustworthy AI\u2019[4]<\/sup><\/a> , and the European Union\u2019s AI Governance Regulation1689\/2024[5]<\/sup><\/a>. The EU Regulation is relied upon as the primary benchmark for measurement and assessment because it is the most comprehensive and binding international framework for the regulation of AI.<\/p>\n

This framework is also characterised by its emphasis on the centrality of fundamental freedoms and democracy, including the protection of personal data, privacy and digital security for users, and ensuring they are fully informed about how their data is collected and used. Furthermore, the EU\u2019s regulatory frameworks place significant and detailed emphasis on a risk-assessment approach in the regulatory documents following the AI Act, which have categorised the types of risks anticipated from the use of AI, in which areas, and classifying the severity of these risks \u2013 ranging from those that are intolerable to those falling into the high, medium and low categories \u2013 whilst setting out the legal consequences of each, providing real-world examples of risks, and specifying unacceptable practices on the part of service providers. [6]<\/a><\/sup><\/p>\n

 <\/p>\n

First: AI Governance: Technical Specifics, Prioritisation Challenges and Regulation Around the World<\/strong><\/span><\/h4>\n

The governance of AI poses a major challenge to national systems around the world, given the rapid pace of development of AI technologies \u2014 which is far outstripping legal and regulatory frameworks \u2014 as well as the varying levels of understanding amongst policymakers and decision-makers worldwide regarding how this technology works. Although regulating this technology is difficult, the risks associated with it necessitate regulation to mitigate those risks. One of the greatest challenges associated with regulating AI lies in clearly and definitively defining the term, which is a major point of contention,[7]<\/sup><\/a> ; however, what the various definitions have in common is that it is the process that makes machines more intelligent. The difficulty in defining it stems in part from the widespread confusion between AI on the one hand, and algorithms on the other; whilst AI is, in fact, merely a branch of algorithms.[8]<\/sup><\/a><\/p>\n

AI governance encompasses a wide range of issues that require discussion, regulation and codification, whether in the form of laws or guiding frameworks; These include ethical considerations and practices that must be taken into account during regulation, as well as administrative and regulatory parameters concerning the actors involved in AI technologies, who are also part of the AI value chain \u2013 such as technology developers, governments and individual consumers\u2014 the limits of their responsibilities, the distribution of burdens amongst them, the nature of the sectors into which AI will be integrated, the resources required to implement these technologies, how their operation will be reviewed and monitored, and the criteria for determining the success of the technology\u2019s adoption.<\/p>\n

Although there are numerous issues that must be addressed and positions taken on when considering an integrated framework for the governance of AI, as outlined above, it is notable that ethical standards and the management of risks arising from the use of AI dominate the discussions. This is due to the prevailing fears and doubts amongst the public and those concerned with issues of rights and freedoms regarding the consequences of adopting AI in an unregulated and irresponsible manner, which may lead to the violation of the personal rights of individuals and groups, as well as threatening their economic rights, in addition to the significant economic and financial losses that may result from adopting this technology without the capacity to address its errors and contain the associated risks.[9]<\/sup><\/a> These concerns are inextricably linked to issues of trust in the governments and entities responsible for implementing AI applications, which further exacerbate these fears.<\/p>\n

Attempts to govern AI vary amongst the initiatives put forward by governments around the world, in what are known as national AI governance strategies, which specifically set out the objectives of adopting such strategies in a given country, as well as the expected commitments,[10]<\/sup><\/a> Other regulatory efforts take the form of laws and legislation governing AI, as well as guidelines and recommendations from international organisations such as UNESCO and the OECD, designed to assist governments in finding the optimal approach to governance. In addition, public policy and legal experts are proposing governance approaches aimed at maximising benefits and minimising risks.<\/p>\n

The fact that these frameworks are mentioned does not imply that they are entirely separate; rather, this review aims to highlight the technical complexity and the current state of attempts to regulate AI, which are still in a phase of trial and error; It also aims to highlight the gap between the overarching objectives and the principles that must be observed (such as non-bias, digital security and privacy), on the one hand, and the difficulty of technically implementing these principles, on the other.<\/p>\n

Various approaches to the governance of AI have been put forward by public policy experts and legal scholars, and these vary according to the specific constraints and priorities of each system. For example, approaches and systems of adaptive AI governance have emerged[11]<\/sup><\/a>, which are centred on the principle of not restricting creativity and innovation in AI technology. This is achieved by ensuring flexibility and a rapid response to changes and emergencies that arise during its use, without being constrained by the need for prior orders for intervention and remediation from official regulatory bodies.<\/p>\n

There is also an approach that runs counter to adaptive governance: centralised governance systems, which are underpinned by a fear that companies will dominate the management and direction of technology in a way that primarily reflects their own interests. Consequently, this system seeks to ensure that governments centrally manage and steer AI technology, thereby preventing corporate misdirection that could harm the interests of consumer groups. A middle ground between the two systems is what is known as hybrid\/decentralised AI governance[12]<\/sup><\/a> , which seeks to ensure flexibility, allowing companies and AI developers to develop the technology and make continuous adjustments to it without the need for complex prior approvals, but in coordination with governments to keep them informed and to ensure transparency and accountability. This debate on governance systems is accompanied by calls to adopt an inclusive, participatory approach to AI governance, with the aim of reaching a societal consensus on which AI applications should be adopted, what the expected benefits and outcomes are, and how the burdens should be distributed amongst the various stakeholders.[13]<\/sup><\/a><\/p>\n

Compounding the difficulty of developing governance approaches for AI are the differing national priorities of governments, which drive their adoption of AI depending on the sectors in which they wish to apply the technology, or in accordance with specific priorities and values. For example, some countries seek to gain a competitive advantage in specific areas by adopting AI technologies, such as in the armaments, defence and security sectors, or in other public services such as the health sector, particularly in countries facing labour shortages and ageing populations.<\/p>\n

Governments\u2019 value-based priorities also influence approaches to AI governance. Take the European Union, for example, which has framed its 2024 AI Regulation[14]<\/sup><\/a> around priorities such as the protection of personal data, the prevention of bias, and risk management \u2013 with all that this entails in terms of restrictions on AI providers and developers.[15]<\/sup><\/a> The EU\u2019s regulation of AI is characterised by its reliance on a pre-existing legislative framework, which sets out the ethical parameters for the operation of AI, such as the General Data Protection Regulation, which was enacted in 2016 and sets out data protection requirements that must be observed in AI applications, particularly in terms of fairness, transparency, accountability and privacy. Therefore, in practice \u2013 even though it does not explicitly mention AI \u2013 the law imposes restrictions on the use of AI to categorise individuals through what is known as \u2018profiling\u2019 based on their personal data, which paves the way for tracking them and making predictions about their future actions and behaviour. This could put them at risk of personal harm or discrimination, or lead to AI applications making formal decisions accordingly, whether in the form of recommendations or final decisions.<\/p>\n

These restrictions include regulatory and classification restrictions on personal data used during the database construction phases and the manner in which it is processed, depending on its level of sensitivity, as well as time limits on the maximum period for which individuals\u2019 personal data may be retained.[16]<\/sup><\/a> The AI Act has built upon this with detailed documents issued following its enactment, requiring applicants to disclose and publish the nature of the data on which they relied when training AI models, in both a technical format and a narrative format that is easy for the public to understand and monitor.[17]<\/sup><\/a> It also requires the documentation of all stages of data use by service developers and providers, thereby enabling accountability and audit. Furthermore, the European AI Act ensures that its provisions and subsequent implementing rules comply with the Charter of Fundamental Rights of the European Union.[18]<\/sup><\/a><\/p>\n

 <\/p>\n

In contrast, the US tends towards a model of voluntary guidelines,[19]<\/sup><\/a> , which implicitly falls under the category of \u2018soft law approaches\u2019\u2014that is, non-binding techniques and rules.[20]<\/sup><\/a> This latter category includes calls for companies working in the field of AI to develop a set of guidelines autonomously and voluntarily, as a commitment to uphold these values, rather than resorting to formal legal regulation with all the burdens and restrictions it entails, which would slow down the development of AI technologies.<\/p>\n

The choice of an AI governance and management framework is also linked to the type of technology and applications to be adopted from the AI matrix, each of which has different management requirements. There is more than one type of AI, depending on its capabilities and the nature of the tasks it is expected to perform. In terms of capabilities, there is narrow or weak AI (Artificial Narrow\/Weak Intelligence), which refers to applications that are well-known amongst consumers \u2013 such as ChatGPT and OpenAI \u2013 which have a significant capacity to perform image and text recognition tasks. This is utilised to automate many repetitive tasks, meaning its outputs are predictable and known in advance; consequently, it cannot go beyond the tasks specified for it.<\/p>\n

Conversely, two other forms of AI are expected to develop: Artificial General Intelligence (AGI) and Super AI. Concerns are growing about these two forms \u2013 despite claims that they remain theoretical for the time being and have no practical applications \u2013[21]<\/sup><\/a> given their ability to mimic human intelligence. These systems are not limited to carrying out specific commands issued to them, but are capable of complex thought processes such as deduction, inference and reasoning once trained, coupled with their ability to scan vast and varied databases in a short space of time \u2013 something humans cannot do \u2013 as well as their capacity for self-learning and for continuously adapting their modes of operation and thinking independently of their developers and manufacturers. This makes the outcomes of their work unpredictable and rapidly changing, making them difficult to control and monitor. General AI also encompasses a model known as neural networks.<\/p>\n

As for its functions, AI is classified according to its purpose and its ability to interact with its surroundings; its sub-types differ depending on how it processes information, how it learns from data, and how it responds to stimuli. The forms of AI vary according to the technology on which they are based, and include types such as:<\/p>\n