In the context of an ever-changing digital panorama, the pervasiveness of Artificial Intelligence (AI) is rapidly transforming various aspects of people's daily lives, including those involving children and teenagers. An important issue, that needs to be addressed, is related to the introduction and enforcement of specific rules to ensure children's rights and special needs in those AI systems created for, accessed by, and impacting minors. Looking at the Italian debates on AI systems, currently, there is no suficient attention to these aspects. The situation is even worse for more general AI-related issues like the creation of a national sandbox by 2026, as expected by the EU AI Act. Accordingly, by briefly presenting the ongoing discussion at the international and national level, this paper would like to set the ground for an Italian debate on how to start thinking and developing a sandbox that considers the special needs and characteristics of children. The Italian scientific community is called by the Italian Strategy for AI [1] to join forces and promote and develop interdisciplinary research projects for social well-being. Accordingly, this paper would like to be a call to not miss the opportunity to frame, from the very beginning, the national sandbox so as to be compliant with children's rights standards and obligations.
The digital panorama is constantly changing, and with it, the way children interact with technology. Artificial intelligence (AI) is rapidly transforming various aspects of our lives, and its reach now extends to applications designed specifically for minors. While AI ofers immense potential for education, entertainment, and personalized learning experiences for children, it also raises some critical concerns regarding safety, privacy, and ethical considerations. Indeed, these systems can significantly afect children’s well-being and day-to-day lives to varying extents. This makes it increasingly crucial to evaluate the current state of this interaction and anticipate future trends to establish ethical and responsible frameworks for AI development.
Countries and international bodies are working on new institutions and portfolios, on the expansion
of regulatory powers, and on new tools to address risks related to digital transformation, including
those related to human rights [
While States have the duty, under international human rights law, to protect people in their jurisdiction
or/and their territory from human rights abuses, corporate responsibility to respect human rights exists
wherever they operate and regardless of their size and industry [
The exercise and protection of human rights, and, thus, of children’s rights, can be afected by "the
way in which digital technologies are designed, developed and deployed, and, therefore it is important
to embed human rights in all the stages of the innovation process, "from research and development to
deployment", and across "a continuum of policy options, from dedicated regulation through collaborative
governance eforts at the global level, soft law instruments, to codes of conduct" (p.53, [
In this work, we intend to briefly present considerations on the Italian debate, framed as well in the international context, on AI systems and regulatory sandboxes for children’s special needs and rights. Therefore, we are going to concisely dive into the regulatory base of this discussion, i.e. the European Union AI Act, and into some useful definitions for the scope of this paper. Our work addresses the following Research Questions (RQs): • RQ1: What is the nature of the discussion in Italy regarding the design, development, and deployment of AI systems that comply with children’s rights standards and obligations, and which specific topics does it address? • RQ2: What initiatives exist in Italy aimed at establishing a regulatory sandbox? We will present some main directions that are being undertaken at the national level to see where we are going, and where we should go, in order to both design, develop and deploy AI systems and establish national regulatory sandboxes that take into account children’s rights and special needs.
Accordingly, we will divide Section 4 into two subsections: one aiming at highlighting the existing discussion, at the national level, on how to design, develop and deploy AI systems in compliance with children’s rights standards and obligations, and one highlighting the initiatives for establishing a national regulatory sandbox comprehensive of a children’s rights point of view. Each subsection will briefly present both the international discussions, to the extent it is part of the frame that can shape the Italian debate on such issues, and the Italian debate on the topic, from which final considerations and possible future developments will be drawn.
In particular, to analyze the Italian debate on this topic, we consulted the web pages 1 of the following
relevant entities:
1. National Cybersecurity Agency (ACN), the agency responsible for safeguarding national
interests, security and resilience in cyberspace [
Given also the indication provided by the EU AI Act of creating at least one national regulatory
sandbox by August 2026 (Art.57, [
On March 2024, the United Nations General Assembly adopted a resolution, "Seizing the opportunities
of safe, secure and trustworthy artificial intelligence systems for sustainable development" , emphasizing
that "human rights and fundamental freedoms must be respected, protected and promoted throughout
the life cycle of artificial intelligence systems" , and calling on the Member States and stakeholders to
"refrain from or cease the use of artificial intelligence systems that are impossible to operate in compliance
with international human rights law or that pose undue risks to the enjoyment of human rights, especially
of those who are in vulnerable situations", and to facilitate "the development and implementation of
efective, internationally interoperable frameworks, practices and standards for training and testing artificial
intelligence systems (...) to help protect individuals (...) throughout the life cycle of artificial intelligence
systems (par. 5 and 6 letter h, pp.5-6 [
A few months later, in August 2024, the first legal framework on AI, the EU AI Act [
The Act addresses AI-related risks by dividing them into four levels: "minimal risk", "limited risk",
"high risk", and "unacceptable risk"[
Article 57, for example, establishes the creation of at least one national regulatory sandbox to be
operational by August 2026 [
About a year before the AI Act came into force, on June 2023, during the time of the inter-institutional negotiations, over 2000 children’s rights, parents’ and mental-health stakeholder organizations, through a coalition, called on negotiators not to miss the occasion of the AI Act to protect children, and urged to recognize their specific rights, needs and vulnerabilities, to ensure safer AI for children [17].
Accordingly, recognizing and highlighting that children have specific rights as enshrined in the
United Nations Convention on the Rights of the Child (UNCRC) and as further implemented by the
UN General Comment No.25 on the application of these rights in relation to the digital environment,
the AI Act foresees a "Fundamental rights impact assessment for high-risk AI systems" in its Article 27
[
These are important steps toward ensuring more responsible and safe AI systems for children. Nevertheless, the discourse concerning children and AI should not be limited to their safety and protection, nor solely to the development and pre-deployment phase of an AI system designed for, accessed by or impacting children.
On 5 September 2024, the Council of Europe opens for signature its "Framework Convention on Artificial Intelligence and human rights, democracy and the rule of law" , the first international legally binding instrument, aiming to "ensure that activities within the lifecycle of artificial intelligence systems are fully consistent with human rights, democracy and the rule of law" (Art.1(1), [18]). Mindful of the UNCRC, the Council’s Framework Convention requires State Parties to "take due account of any specific needs and vulnerabilities in relation to respect for the rights of persons with disabilities and of children" (Art.18, [18]).
Therefore, the debate should also revolve around how to consider, include, promote and uphold children’s rights, and how to do that from the beginning of an AI system’s lifecycle. This is essential also to avoid dificult, expensive and time-consuming attempts to retrofit AI systems already developed by trying to make them in compliance with children’s rights and international human rights laws. Moreover, it will make it easier, for the diferent stakeholders involved, to identify, report and address risks and challenges from the very beginning of the system’s existence. Accordingly, similarly to the concept of "security-by-design" and "privacy-by-design", this paper advocates for a children’s rights-bydesign approach from the first to the last phase of the life of an AI system (for the concept’s definition, please refer to the Appendix 6). Indeed, the "by-design thinking" (p.4, [19]), has traditionally been applied in the area of privacy, data protection, and security, but it has begun to spread also throughout the legal system [19].
Keeping in mind this whole context, in the following Section we will provide a brief overview of the International and Italian debate on applying these concepts during all the phases of an AI system’s lifecycle and on the creation of a regulatory sandbox that considers and is compliant with children’s rights standards and obligations.
According to the AI Act, the regulatory sandboxes ofer an environment to "develop, train, validate and
test" [
Accordingly, since this paper would like to be a call to start a national discussion on how to achieve this goal, and on how to create and make operational a regulatory sandbox that not only identifies and assesses risks but also considers children’s specific rights and needs, in this Section we will provide a brief overview of the existing debates, at the international and national level, on both a children’s rights approach to AI systems and on regulatory sandbox (supposed to be created by 2026).
The relationship between children and AI is not easy. UNICEF [20] acknowledges that the interaction between children and AI systems is not limited only to those systems used or designed for children, but also to those systems not meant for them but with which they may interact, or systems that can directly or indirectly afect them [20]. However, considering and including children’s rights during the whole phases of a system’s lifecycle is a challenging and delicate task.
Nevertheless, there have been some policy initiatives, at the International and European level, on AI systems children’s rights.
In 2024, the Council of Europe released a study "on the rights of the child and artificial intelligence , mapping recent "legal frameworks that address AI in the context of children’s rights" [21]. In this overview, are also reported some frameworks that "do not specifically focus on AI and children’s rights but are still relevant to consider as they relate to the protection and empowerment of children in today’s digital world" [21].
Among all these initiatives, the ones from UNICEF [20], World Economic Forum (WEF) [22], and the Institute of Electrical and Electronics Engineers Standards Association (IEEE SA) [23] are those that explicitly consider providing practitioners and developers, working on AI systems for, accessed by and/or impacting children, with guidance on how to take into account and operationalized, during their daily work’s duties, children’s rights in all the phases of a system’s lifecycle.
Indeed, UNICEF Policy Guidance on AI for Children [20], is an attempt to connect government policies with the private sector’s practices. Therefore, the document provides nine requirements for child-centered AI, and, to support their implementation, a list of online resources and a set of practical implementation tools for both policymakers and developers [20]. The WEF’s Artificial Intelligence for Children Toolkit [22], with its FIRST approach (Fair, Inclusive, Responsible, Safe, and Transparent), is also addressed to members of a product team. It aims to help "make sure that AI respects the rights of children and has a positive impact in their lives"[22]. Finally, the contribution from IEEE SA on the Age Appropriate Digital Services Framework Based on the 5Rights Principles for Children[23], although not specific on AI, establishes a set of processes by which organisations seek to make their services age appropriate. This Framework[23], based on the work of the 5Rights Foundation, has been recognised in 2023 to serve as the foundation for a new Workshop Agreement with the European Committee for Standardization (CEN) and European Electrotechnical Committee for Standardization (CENELEC) [24]. For a more in-depth and multidimensional analysis of these three documents, please refer to [25], while, for an overview and brief description of all the international available guidelines highlighted by the Council of Europe, please refer to Table 1 available in the Appendix.
To map any existing similar discussion at the Italian level, and to answer the first research question (RQ1: What is the nature of the discussion in Italy regarding the design, development, and deployment of AI systems that comply with children’s rights standards and obligations, and which specific topics does it address?), the websites of the most important national agencies, authorities, and non-governmental and non-profit organizations were checked to find any information publically available on how to design, develop and deploy AI systems in line with children’s rights standards and obligations.
The National Cybersecurity Agency (ACN), as the agency responsible for safeguarding national interests, security and resilience in cyberspace, has joined the "Guidelines for secure AI system development"[26], promoted by the National CyberSecurity Centre of the United Kingdom and published in November 2023. The Guidelines are intended for "providers of any systems that use artificial intelligence (AI), whether those systems have been created from scratch or built on top of tools and services provided by others" [26], and are divided into four key areas ("secure design, secure development, secure deployment, and secure operation and maintenance") to set security as "core requirement" throughout all the system’s life cycle [26]. Although these guidelines are intended to be read and followed by diferent stakeholders, from developers to managers and decision-makers, they do not ofer any further specific considerations for AI systems intended for, accessed by or impacting children.
Nevertheless, in the "Cybersecurity Research and Innovation Agenda (2023-2026)" [27], the ACN, on the research areas No.5 "Societal Aspects", calls for an interdisciplinary approach to topics such as "redefinition of boundaries of human-computer interaction and related security risks" (topic 5.1.1, p.23 [27]), or "rules and ethical principles on safe cyberspace" (topic 5.3.1., p, 25 [27]). Moreover, the "National Cybersecurity Strategy (2022-2026) [28] suggests, in the context of the fight against cybercrime, to also focus on the protection of minors [28]. And in the Strategy’s "Implementation plan" [29], with Measure No.73, the ACN foresees to "prepare and implement an autonomous national strategy, with a related action plan, dedicated to the online protection of minors from cybercrime (...)" (p.20, [29]).
From the Agency for the Italia Digitale (AgID)’s website, searching under the Artificial
Intelligence’s area of intervention, it is possible to access 2 the "Italian Strategy for Artificial Intelligence
2024-2026" [
On the other hand, the Italian Data Protection Authority, as an independent authority entitled to the protection of fundamental rights and freedom relating to the processing of personal data [30], provides an information web page on children, new technology and data protection, and provides people (including children and youth) with various vademecum, such as on advice to adults for children’s safe use of new technologies [31] and smart toys [32].
In the context of minors, new technology and data protection, the Authority also recalls the "Guidelines on Transparency under Regulation 2016/679" [33], which provide "practical guidance and interpretative assistance (...) on the new obligation of transparency concerning the processing of personal data under the General Data Protection Regulation" [33]. In particular, the document defines specific guidance when providing information to children (p.10, [33]). When a "data controller is targeting children", for example, should be aware that "their goods/services are particularly utilised by children (...), it should ensure that the vocabulary, tone and style of the language used is appropriate to and resonates with children (...)", and provide an example of a child-centred language that can be used alternatively to the original legal one [33]. Therefore, these guidelines, providing practical guidance and interpretative assistance, can give, not only to "data controller" but also to developers, more detailed ideas of children’s special needs, how to translate and incorporate them into their work, and how to practically address issues that may rise.
In October 2024, the Italian Data Protection Authority, in a roundtable with the Data Protection and Privacy Authorities of the States belonging to the G7, agreed on a statement on AI and children [34]. The involved authorities highlighted concerns about "potential violations of privacy and data protection linked to the use of AI systems which could have serious implications for children and young people" [34]. Accordingly, the authorities recognised the need for "AI systems’ developers and users to account for specific age-appropriate measures to allow children and young people to safely use AI-enabled technologies" [34]. Such measures should ensure, among other things, a design that supports children’s best interests and adopt a privacy-by-design principle.
The Italian Authority for Childhood and Adolescence also focuses on children, digital space and AI. In its "Report to the Parliament 2023" [35], the Authority also refers to media, digital space and children’s rights and the related activities conducted in 2023, from a project culminating with a "Children’s Manifesto on Rights in the Digital Environment", where children identify 10 principles to take into account when minors deal with digital space, to a training day during the Safer Internet Day on topics such as children’s safety and security online and AI [35]. Moreover, over the years, it has also worked on various publications on such topics through comics (see [36]) and translation (also in a child-friendly format) of international relevant documents (see [37]).
In addition, the Authority established a Memorandum of Understanding with both Save the Children Italia and Fondazione SOS Telefono Azzurro. Indeed, Save the Children is also active in this topic through projects and reports. Such as DIG4Future - Digital competencies, Inclusion and Growth for Future Generations [38], which aims to foster youth’s digital competencies and promote responsible use of technology by training teachers and co-creating instruments and inclusive educational paths about AI, and the Atlante dell’infanzia (a rishio) in Italia 2023: Tempi Digitali [39], which analyses new scenarios, situations, risks and opportunities, about children, their needs and desires, coming from the technology and its spread.
Telefono Azzurro has also been active on the topic through many dossiers. The 2020 investigation
Together for a Better Internet [40], conducted together with DoxaKids, for example, provides an overview of risks and children’s needs in relation to the web and highlights the contribution that AI can bring to achieve a more safe and secure web for children (i.e.: the hashing technology to help identify and takedown from the web images of sexual abuses and exploitation of children) [40]. Moreover, the dossier is making a "call to action" that encompasses investing in "multi-sector research" on risks and opportunities to achieve a greater understanding and develop concrete actions, and involving technology companies and leaders to really implement solutions for a more secure and child-centred web (p.42, [40]).
Several European states include the use of sandboxes as a means to build a comprehensive legal
framework for artificial intelligence. This trend is supported by the EU, which views regulatory
sandboxes as facilitators of innovation and recognizes them as an important tool in future regulatory
activities concerning AI. A regulatory intervention for the definition of this tool was provided by the AI
Act, definitively approved on May 21st 2024, which gives a definition of AI sandboxes in Article 57 [
• Analysis. Understanding the impact of AI on children within national and regional contexts.
The sandbox applies relevant policies, incorporates children’s inputs, and fosters co-design with stakeholders to address the social, technological, and cultural aspects of AI’s impact. • Design. Testing strategies for integrating child-centered AI policies into products and services.
Ensuring transparency through open evaluation and public scrutiny of projects. • Implementation. Support iterative testing and adaptation of designs while aligning implementation decisions with legal and policy guidelines. • Evaluation. Verify compliance with child-centered AI requirements, monitor impacts using metrics, and refine policies and strategies based on data-driven insights.
The proposed analysis is very interesting as it involves various subjects (both institutional and noninstitutional) and each of them has an important role in evaluating and determining what impact an artificial intelligence system can have on minors.
Among the institutional subjects that should be involved in the implementation of a regulatory sandbox for minors in Italy, there is the Italian Authority for Childhood and Adolescence, which recently has expressed its opinion regarding the worthy legislative bills no. 1136, 1160 and 1166 (Protection of minors in the digital dimension) [42]. The analysis of the information published on the webpages of the competent Italian bodies, as listed in the Methodology Section, has highlighted how the issue of regulatory sandboxes for minors in Italy has not yet been properly addressed and taken into consideration by the current legislation nor by any of the above-mentioned bodies.
At first glance, to answer RQ2 ( What initiatives exist in Italy aimed at establishing a regulatory sandbox?) we can highlight some important interventions in this field. First of all, there is a draft of law (DDL 1066) [43] which, in its article 4, provides for the establishment of experimentation spaces relating to the use of artificial intelligence systems aimed at pursuing the innovation of services and products in various sectors including education, where children are more involved. New rules for minors’ access to dangerous content on the web have been introduced in Italy by the Autorità per le Garanzie nelle Comunicazioni (AGCOM) with Resolution 9/23/Cons [44] in force since November 21st, 2023. This measure requires telephone operators to implement a control system on SIM cards registered to minors, which automatically blocks access to inappropriate content. For adults, the filter is activated only manually. The Italian Communications Authority regulates and supervises the sectors of electronic, audiovisual, editorial, postal communications and more recently also online platforms, as specified on its website. The aim of this measure is to limit the risks deriving from the indiscriminate use of the web by young people, restricting the possibility of browsing without control. Even though many telephone companies already ofer parental control systems, these are often associated with paid services. The AGCOM resolution mandates that these controls be provided free of charge. Operators must implement automatic filtering for eight categories of high-risk content, including adult material, gambling, weapons promotion, incitement to violence or self-harm, hatred or discrimination, harmful health practices, sect-related content, and tools for untraceable online activity. This list was determined through a public consultation that involved trade associations in the communications sector, consumer associations and individual operators. However, the categories are generic, and it will be up to the telephone operators to identify and communicate the precise list of sites prohibited for minors. In March 2024, a bill [43] regarding "Regulations for the development and adoption of artificial intelligence technologies" was presented to the Senate of the Italian Republic. The text emphasizes that artificial intelligence opens new opportunities, such as increased productivity, which could result in greater value-added production within the same amount of time or a reduction in working hours while maintaining the same value, thereby improving workers’ quality of life. However, AI also presents ethical and social risks. For innovation to become a driver of progress, responsible development is crucial. Transparency in data management, the reliability of technologies, the protection and privacy of systems, the impartiality of outputs, and the accessibility of services are all fundamental aspects of the implementation of ethical AI.
Up to now, the discussion on children’s rights and AI systems, from a policy point of view, has mainly been focused on the principle of protection, by fostering safety measures, risks and impact assessments and a privacy-by-design approach. This is particularly true at the International and European levels, even if some initiatives aimed at closing the distance between policy guidelines on children and AI and the technical side of innovation have been attempted by suggesting guidelines or frameworks addressed to developers on how to design, develop and deploy AI systems compliant with children’s rights (see, for example, [20] and [23]).
In Italy, this discussion, on providing practitioners working on AI systems with guidelines to follow so as to adopt a children’s rights approach to the design, development and deployment of AI systems for, accessed by and/or impacting children, is still in its infancy. Consequently, the risk is that the AI systems developed and tested, in the regulatory sandbox that has to be established by 2026, are likely not to be fully compliant with children’s rights standards and obligations so as to protect children while also ensuring their provision and participation rights.
Adopting a children’s rights-by-design approach that, similar to the idea behind concepts like "privacy by design" and "security by design", would mean giving due weight and including children’s rights consideration in every phase of the life of an AI system, particularly but not solely for those systems targeting or accessed by children. This approach could lead to fewer risks assessed and verified during the "fundamental rights impact assessment", since the system is designed to be compliant from the beginning of its development, and in the regulatory sandboxes in the pre-deployment phase, since risks assessment and mitigation strategy were already identified and addressed during the previous phases of the AI system life cycle.
We believe that this will not only be less time-consuming and fasten the process, but will also be less expensive for companies. Moreover, it will make it easier for stakeholders and experts to identify all risks and properly address them during the process, instead of leaving them to manage everything in the last phases, risking leaving some threats or challenges unseen.
The Italian scientific community is called by the Italian Strategy for AI [
Article 57 of the AI Act establishes that each Member State must establish at least one regulatory
testing space for AI at the national level and that this space must be operational by August 2026 [
A. Definitions
• Children’s Rights-by-Design. The concept was already presented by the Digital Futures
Commission and 5 Rights Foundation in a Report published in 2023 [45]. Acknowledging that designing
for children’s rights is not easy, but that retrofitting design to respect them after a product has
already been developed can be both dificult and expensive, the report provides 11 principles
grounded in the UN Convention on the Rights of the Child (UNCRC) ("equity and diversity",
"best interest", "consultation", "age appropriate", "responsible", "participation", "privacy", "safety",
"wellbeing", "development", and "agency") to "help realise children’s rights when designing digital
products and services", and to be applied in all phases of an innovation process (that they divide
in discover, define , develop, and deliver) [45]. Adopting a children’s rights-by-design approach
would mean giving due weight and including children’s rights considerations in every phase of
the life of an AI system, particularly but not solely for those systems targeting or accessed by
children. This approach, which should be comprehensive in terms of both privacy and security
by design approach, should also be applied together with an ethical and societal evaluation;
• Regulatory Sandboxes. Article 3(55) AI Act [
The term regulatory sandbox covers a wide variety of programs run by national financial regulators in order to allow for controlled testing by private firms of innovative financial products and services. In general, sandboxes provide a ‘safe space’ for fintechs and financial firms to ofer real products to real customers with the benefit of a waiver, or a significant relaxation of otherwise applicable regulations [46].They are typically justified as a means of supporting consumer-benefitting financial innovation, facilitating financial inclusion, improving the eficiency and competitiveness of domestic financial institutions, and advancing regulators’ own understanding of the emerging innovative technologies [47]. While described as a “laboratory environment”, its primary function is to test innovations within an existing regulatory framework through a collaborative process involving both the regulator and participating businesses. Like its technological counterpart, the regulatory sandbox aims to mitigate risks. However, the type of risk involved difers significantly from that of computer systems, requiring a distinct and adaptive approach to designing diferent sandboxes based on the involved participants [48].
This study has been partially supported by the following projects: SERICS (PE00000014) under the MUR National Recovery and Resilience Plan funded by the European Union – NextGenerationEU.
The authors have not employed any Generative AI tools. [17] 5Rights Foundation, COST Action into Problematic Usage of the Internet, ECPAT International, Eurochild, European Parents’ Association, Mental Health Europe, Missing Children Europe, and Terre des Hommes, AI Act trilogues: the EU’s last chance to protect children, 2023. URL: https://5rightsfoundation.com/wp-content/uploads/2024/09/Coalition-statement_ Childrens-rights-in-AI-Act-Trilogues.pdf, (Last access online: 02.12.2024). [18] Council of Europe, Council of Europe Framework Convention on Artificial Intelligence and Human
Rights, Democracy and the Rule of Law, 2024. [19] C. Djefal, Children’s rights by design and internet governance: Revisiting general comment no.
25 (2021) on children’s rights in relation to the digital environment, Laws 11 (2022) 84. [20] UNICEF, Policy guidance on AI for children, 2021. URL: https://www.unicef.org/innocenti/reports/ policy-guidance-ai-children, (Last access online: 02.10.2024). [21] Council of Europe, The Alan Turing Institute, Mapping Study on the Rights of the Child and Artificial Intelligence. Legal Frameworks that Address AI in the Context of Children’s Rights, 2024.
URL: https://www.coe.int/en/web/children/publications, (Last access online: 06.12.2024). [22] World Economic Forum, Artificial Intelligence for children. TOOLKIT, 2022. URL: https://www.
weforum.org/publications/artificial-intelligence-for-children/, (Last access online: 05.09.2024). [23] IEEE Standards Association (IEEE SA) , Age appropriate digital services framework, 2021. URL: https://standards.ieee.org/ieee/2089/7633/, (Last access online: 06.12.2024). [24] IEEE, CEN and CENELEC, Age appropriate digital services framework. Workshop Agreement CWA 18016, 2023. URL: https://standards.ieee.org/news/ieee-2089-european-reference-document/, (Last access online: 02.10.2024). [25] D. Caivano, B. F. Nieto, D. Gigante, A. Ragone, S. Tibidò, Ensuring child rights in the age of ai: A multidimensional analysis of existing frameworks, in: Proceedings of the 2024 International Conference on Information Technology for Social Good, 2024, pp. 76–83. [26] Agenzia per la Cybersicurezza Nazionale, Guidelines for secure AI system development, 2023.
URL: https://www.acn.gov.it/portale/linee-guida-ia, (Last access online: 11.12.2024). [27] Agenzia per la Cybersicurezza Nazionale, Cybersecurity Research and Innovation Agenda (20232026), 2023. URL: https://www.acn.gov.it/portale/agenda-di-ricerca-e-innovazione, (Last access online: 11.12.2024). [28] Agenzia per la Cybersicurezza Nazionale, Strategia Nazionale di Cybersicurezza (2022-2026), 2022.
URL: https://www.acn.gov.it/portale/strategia-nazionale-di-cybersicurezza, (Last access online: 11.12.2024). [29] Agenzia per la Cybersicurezza Nazionale, Piano di Implementazione Strategia Nazionale di Cybersicurezza (2022-2026), 2022. URL: https://www.acn.gov.it/portale/ strategia-nazionale-di-cybersicurezza, (Last access online: 11.12.2024). [30] Garante per la Protezione dei Dati Personali GPDP, The Italian Data Protection Authority: Who We Are, 2024. URL: https://www.garanteprivacy.it/web/garante-privacy-en/ the-italian-data-protection-authority-who-we-are, (Last access online: 08.12.2024). [31] Garante per la Protezione dei Dati Personali GPDP, Minori e Nuove Tecnologie - Consigli ai «GRANDI» per un utilizzo sicuro da parte dei «PICCOLI», 2021. URL: https://www.garanteprivacy. it/web/guest/home/docweb/-/docweb-display/docweb/9537410, (Last access online: 07.12.2024). [32] Garante per la Protezione dei Dati Personali GPDP, Smart Toys: i suggerimenti del Garante per giochi a prova di privacy, 2024. URL: https://www.garanteprivacy.it/temi/iot/smarttoys, (Last access online: 08.12.2024). [33] Working Party on the Protection of Individuals with Regard to the Processing of Personal Data.
Article 29 Working Party (WP29) , Guidelines on Transparency under Regulation 2016/679, 2018.
URL: https://ec.europa.eu/newsroom/article29/items/622227, (Last access online: 11.12.2024). [34] G7 Data Protection and Privacy Authorities, Statement on AI and Children, 2024. URL: https: //www.garanteprivacy.it/g7garantiprivacy, (Last access online: 08.12.2024). [35] Autorità garante per l’infanzia e l’adolescenza, Relazione al Parlamento 2023, 2024. URL: https: //www.garanteinfanzia.org/relazioni-annuali, (Last access online: 11.12.2024). [36] Autorità garante per l’infanzia e l’adolescenza, Geronimo Stilton. Alla scoperta del mondo digitale, 2022. URL: https://www.garanteinfanzia.org/pubblicazioni, (Last access online: 11.12.2024). [37] Autorità garante per l’infanzia e l’adolescenza, Comitato sui Diritti dell’Infanzia. Commento Generale No.25 Sui diritti dei minorenni in relazione all’ambiente digitale, 2022. URL: https: //www.garanteinfanzia.org/pubblicazioni, (Last access online: 11.12.2024). [38] Save The Children, DIG4Future - Digital competencies, Inclusion and Growth for Future Generations, 2021. URL: https://www.savethechildren.it/cosa-facciamo/progetti/ dig4future-digital-competencies-inclusion-and-growth-future-generations, (Last access online: 12.12.2024). [39] Save the Children Italia, Atlante dell’infanzia (a rishio) in Italia 2023: Tempi Digitali, 2023. URL: https://www.savethechildren.it/cosa-facciamo/pubblicazioni/ 14-atlante-dell-infanzia-a-rischio-tempi-digitali, (Last access online: 11.12.2024). [40] Telefono Azzurro, and DoxaKids, Together for a Better Internet, 2020. URL: https://azzurro.it/ dossier-digitale/, (Last access online: 11.12.2024). [41] V. Charisi, V. Dignum, Operationalizing ai regulatory sandboxes for children’s rights and wellbeing, in: Human-Centered AI, Chapman and Hall/CRC, 2024, pp. 231–249. [42] Autorità garante oer l’infanzia e l’adolescenza, Parere dell’autorità garante ai sensi dell’art. 3, comma 3, della legge 12 luglio 2011, n. 112, 2024. URL: https://www.garanteinfanzia.org/ sites/default/files/2024-07/parere-ddl-1136-1160-1166-tutela-digitale.pdf, (Last access online: 11.12.2024). [43] DDL No.1066, Norme per lo sviluppo e l’adozione di tecnologie di intelligenza artificiale, 2024.
https://www.senato.it/leg/19/BGT/Schede/Ddliter/58066.htm. [44] Autorità per le Garanzie nelle Comunicazioni, Adozione delle linee guida finalizzate all’attuazione dell’articolo 7-bis del decreto-legge 30 aprile 2020, n. 28 in materia di “sistemi di protezione dei minori dai rischi del cyberspazio, 2023. URL: https://www.agcom.it/documentazione/documento?p_p_auth=fLw7zRht&p_p_id=101_ INSTANCE_FnOw5lVOIXoE&p_p_lifecycle=0&p_p_col_id=column-1&p_p_col_count=1&_101_ INSTANCE_FnOw5lVOIXoE_struts_action=/asset_publisher/view_content&_101_INSTANCE_ FnOw5lVOIXoE_assetEntryId=29649001&_101_INSTANCE_FnOw5lVOIXoE_type=document, (Last access online: 12.12.2024). [45] Digital Future Commission and 5 Rights Foundation, Child Rights by Design, 2023. URL: https: //5rightsfoundation.com/resource/child-rights-by-design/, (Last access online: 31.10.2024). [46] S. T. Omarova, Technology v technocracy: Fintech as a regulatory challenge, Journal of Financial
Regulation 6 (2020) 75–124. [47] H. J. Allen, Regulatory sandboxes, Geo. Wash. L. Rev. 87 (2019) 579. [48] K. Yordanova, The shifting sands of regulatory sandboxes for AI, 2019. URL: https://www.law. kuleuven.be/citip/blog/the-shifting-sands-of-regulatory-sandboxes-for-ai/, (Last access online: 25.11.2024). [49] UN Ofice of the High Commissioner for Human Rights, General comment No. 25 (2021) on children’s rights in relation to the digital environment, 2021. URL: https://www.ohchr.org/en/documents/general-comments-and-recommendations/ general-comment-no-25-2021-childrens-rights-relation, (Last access online: 06.12.2024). [50] UN Human Rights Council, Artificial intelligence and privacy, and children’s privacy Report of the Special Rapporteur on the right to privacy, Joseph A. Cannataci, 2021. URL: https://www.ohchr.org/en/documents/thematic-reports/ ahrc4637-artificial-intelligence-and-privacy-and-childrens-privacy, (Last access online: 10.12.2024). [51] Council of Europe, Convention on Cybercrime, 2001. URL: https://www.coe.int/en/web/ cybercrime/the-budapest-convention, (Last access online: 10.12.2024). [52] Council of Europe, Council of Europe Convention on the Protection of Children against Sexual Exploitation and Sexual Abuse (CETS No. 201), 2007. URL: https://www.coe.int/en/web/conventions/ full-list?module=treaty-detail&treatynum=201, (Last access online: 10.12.2024). [53] Council of Europe, Recommendation CM/Rec(2018)7 of the Committee of Ministers to member