{"id":72135,"date":"2026-07-09T10:55:00","date_gmt":"2026-07-09T08:55:00","guid":{"rendered":"https:\/\/www.lexia.it\/2026\/07\/09\/data-technology-insight-july\/"},"modified":"2026-07-10T12:33:48","modified_gmt":"2026-07-10T10:33:48","slug":"data-technology-insight-july","status":"publish","type":"post","link":"https:\/\/www.lexia.it\/de\/2026\/07\/09\/data-technology-insight-july\/","title":{"rendered":"Data & Technology Innovation | July 2026 Insight"},"content":{"rendered":"\n

AI Act and the \u201cOmnibus VII\u201d package: what changes between new deadlines, simplifications, and prohibitions<\/strong><\/h2>\n\n\n\n

Artificial Intelligence<\/em><\/p>\n\n\n\n

On 29 June 2026, the Council of the European Union gave final approval to the new \u201cDigital Omnibus on AI<\/em>\u201d Regulation<\/a>. The measure amends Regulation (EU) 2024\/1689 (the \u201cAI Act<\/strong>\u201d) and other related legislation, with the aim of simplifying the application of artificial intelligence rules and reducing the administrative burden on businesses, providers, and deployers<\/em>, without compromising the protection of fundamental rights.<\/p>\n\n\n\n

One of the main changes concerns the postponement of the deadlines applicable to high-risk AI systems: 2 December 2027 for standalone high-risk systems (non-embedded systems<\/em>); and 2 August 2028 for systems embedded in products. The deadline for the operational launch of regulatory sandboxes<\/em> is also postponed to 2 August 2027.<\/p>\n\n\n\n

The Regulation introduces two new operational prohibitions from 2 December 2026, in addition to those set out in Article 5 of the AI Act: the use of AI to create or manipulate, without consent, realistic intimate images or videos (sexual deepfakes<\/em>), and child sexual abuse material.<\/p>\n\n\n\n

In terms of simplification, the Regulation addresses overlaps between the AI Act and sector-specific legislation (e.g., medical devices and toys): where special legislation already ensures equivalent or higher safeguards, certain obligations under the AI Act may not apply. Regulation (EU) 2023\/1230 on machinery is now included in Annex I to the AI Act, thereby integrating AI-related safety requirements into the machinery regulatory framework.<\/p>\n\n\n\n

The Regulation also addresses the relationship with the GDPR. A new legal basis allows providers of high-risk systems, in exceptional cases and subject to strict safeguards, to process special categories of personal data (Article 9 GDPR) in order to correct algorithmic distortions (bias<\/em>). In addition, it provides for the possibility of coordinating the fundamental rights impact assessment (FRIA) with the data protection impact assessment (DPIA) required under the GDPR, thereby avoiding duplication.<\/p>\n\n\n\n

The Regulation will enter into force on the third day following its publication in the Official Journal of the European Union.<\/p>\n\n\n\n

AGCM opens an investigation into the impossibility of backing up an Apple device to a cloud<\/em> service other than iCloud (breach of the interoperability obligation)<\/strong><\/h2>\n\n\n\n

Digital Markets Act<\/em><\/p>\n\n\n\n

By resolution<\/a> of 9 June 2026, the Italian Competition Authority (Autorit\u00e0 Garante della Concorrenza e del Mercato, \u201cAGCM<\/strong>\u201d) opened an investigation against Apple Inc., Apple Distribution International Ltd, and Apple Italia S.r.l. (\u201cApple<\/strong>\u201d) for an alleged breach of the interoperability obligation under Article 6(7) of Regulation (EU) 2022\/1925 (Digital Markets Act, \u201cDMA<\/strong>\u201d), with reference to the iOS and iPadOS operating systems and cloud services competing with iCloud.<\/p>\n\n\n\n

This is the first time that the AGCM has exercised the investigative powers granted to it under the DMA.<\/a> The proceedings originate from a complaint filed in April 2025 concerning the impossibility for users to perform a full backup<\/em> of their iPhone or iPad on a cloud service other than iCloud. This limitation would allegedly be attributable to a technical choice by Apple, which would prevent competing providers from accessing the components of iOS and iPadOS required to create a complete backup<\/em>, a functionality instead guaranteed to iCloud.<\/p>\n\n\n\n

Apple, which has been designated as a gatekeeper<\/em> under the DMA for iOS and iPadOS, is required to ensure that third-party cloud <\/em>service providers have effective interoperability with the same hardware and software features available to iCloud. If confirmed, the conduct would favour iCloud to the detriment of competitors, discouraging users from relying on alternatives and undermining the contestability of digital markets.<\/p>\n\n\n\n

The Court of Cassation on the validity of a \u201ctick-box\u201d mechanism for online acceptance of unfair terms<\/strong><\/h2>\n\n\n\n

E-commerce<\/em><\/p>\n\n\n\n

By order<\/a> no. 20945 of 20 June 2026 (the \u201cOrder<\/strong>\u201d), the Italian Court of Cassation ruled on the validity of a \u201ctick-box\u201d mechanism for accepting unfair terms in online contracts, finding that it did not comply with the requirement under Article 1341(2) of the Italian Civil Code, read in conjunction with Legislative Decree no. 70\/2003 on e-commerce.<\/p>\n\n\n\n

The Court held that this method was insufficient to ensure specific written approval, which, in distance contracts not subject to a mandatory written form requirement (ad substantiam<\/em>), may take the form of a simple electronic signature (Article 3, point 10, of the eIDAS Regulation).<\/p>\n\n\n\n

The Order does not introduce a new principle, but is consistent with settled case law (including judgment no. 9413\/2021), according to which a simple electronic signature is sufficient for distance contracts, including access to a reserved area and pressing an acceptance button (\u201cpoint and click<\/em>\u201d).<\/p>\n\n\n\n

In the case at hand, the contract had been concluded by ticking the boxes corresponding to the service, and the unfair term had been approved through a double tick-box mechanism. However, the PDF printout filed in court did not prove specific acceptance equivalent to signature of the general contract. The Court clarified that it is sufficient for the service provider to make available on its website a form<\/em> enabling the signatory to specifically approve the unfair term by means of a simple electronic signature, whereas mere \u201cticking\u201d is not sufficient.<\/p>\n\n\n\n

Generative artificial intelligence and copyright protection: the 2026 AGCOM Report<\/strong><\/h2>\n\n\n\n

Copyright<\/em><\/p>\n\n\n\n

On 17 June 2026, the Italian Communications Authority (Autorit\u00e0 per le Garanzie nelle Comunicazioni, \u201cAGCOM\u201d) published its 2026 Artificial Intelligence Report, which outlines the technological, economic, and legal transformations brought about by the development of AI models. The Report is divided into two sections: the first provides a technical and economic analysis of the evolution of Large Language Models<\/em>; the second develops a legal and regulatory analysis of the interaction between the AI Act, the Digital Services Act, and copyright law, with a specific focus on the protection of human creativity.<\/p>\n\n\n\n

The central principle of the Report is the requirement of human creative authorship as a precondition for copyright protection under Article 2575 of the Italian Civil Code and Law no. 633\/1941. The Report critically addresses whether output<\/em> autonomously generated by an AI system may qualify as an \u201cintellectual work<\/em>\u201d and whether the system may be considered an \u201cartificial author<\/em>\u201d, tending to exclude protection in the absence of a recognisable human creative contribution.<\/p>\n\n\n\n

The second key issue concerns the training of generative models and the lawfulness regime applicable to text and data mining<\/em>. The Report highlights that the machine-readable opt-out<\/em> mechanism (Articles 3 and 4 of Directive 2019\/790) is the instrument used to balance the development of models with the right of rightholders to reserve the use of their content. This is complemented by the transparency obligation under Article 53 of the AI Act, which requires providers of general-purpose models to provide a summary of the protected content used during the training phase. <\/p>\n\n\n\n

AGCOM, in its capacity as Digital Services Coordinator, emphasises the value of fair compensation, already tested in relationships between publishers and platforms, as a paradigm for redistributing the value generated by the use of content in the training of generative AI. To this end, the Authority has launched a permanent technical working group on AI, copyright, and pluralism.<\/p>\n\n\n\n

AGCM fines Philip Morris Italia EUR 7 million for advertising claims<\/em> relating to IQOS<\/strong><\/h2>\n\n\n\n

Unfair commercial practices<\/em><\/p>\n\n\n\n

By decision<\/a> of 9 June 2026 (the \u201cDecision<\/strong>\u201d), the Italian Competition Authority (Autorit\u00e0 Garante della Concorrenza e del Mercato, \u201cAGCM<\/strong>\u201d) imposed a fine of EUR 7 million on Philip Morris Italia S.r.l. for an unfair commercial practice. The company used expressions such as \u201csmoke-free\u201d, \u201csmoke-free products\u201d, and \u201cdesigning a smoke-free future\u201d in relation to IQOS heated tobacco products, causing consumers, including minors, to mistakenly believe that such products were free from harmful effects or less harmful than traditional cigarettes.<\/p>\n\n\n\n

The proceedings originated from a report submitted by the Ministry of Health in March 2025 concerning an advertising campaign entitled \u201cIn Italy to build a smoke-free future together<\/em>\u201d. <\/a>AGCM found that the expression \u201csmoke-free\u201d does not merely describe the absence of combustion, but constitutes a misleading message capable of creating in the average consumer the perception that such products are harmless or less harmful than cigarettes.<\/p>\n\n\n\n

The message is misleading because it is not supported by any scientific evidence: the Ministry of Health, the Italian National Institute of Health (Istituto Superiore di Sanit\u00e0), and the WHO have clarified that the aerosol emitted by heated tobacco products falls within the definition of smoke, that such products contain nicotine in concentrations similar to those of traditional cigarettes, and that there is no evidence that they are less harmful. AGCM found that the company was aware of this scientific uncertainty but nevertheless intensified its communication strategy centred on the \u201csmoke-free\u201d claim from 28 June 2022.<\/p>\n\n\n\n

Final version of the Code of Practice on Transparency of AI-Generated Content<\/em> published<\/strong><\/h2>\n\n\n\n

Artificial intelligence<\/em><\/p>\n\n\n\n

On 10 June 2026, the European Commission published the final version of the Code of Practice on Transparency of AI-Generated Content<\/a><\/em> (the \u201cCode<\/strong>\u201d), a voluntary adherence tool that translates the transparency obligations for AI-generated content under Article 50 of the AI Act into operational measures. To be included in the first public list of signatories, providers<\/em> and deployers<\/em> must adhere by 22 July 2026<\/span>.<\/p>\n\n\n\n

The Code confirms its two-section structure (Section 1 for providers<\/em> and Section 2 for deployers<\/em>), introducing certain changes compared with the second version of the Code<\/a>: <\/p>\n\n\n\n