When in Rome, do as the Romans do? Rome court ‘misinterprets’ CJEU YouTube judgment and reduces criteria for primary liability of platform operators to checklist

Michelangelo's Cumaean Sybil in 
the Sistine Chapel

In the ancient Greek and Roman world, sybils were priestesses who acted as gods’ intermediaries with the mortals to anticipate future events. Their prophecies, however, were delivered in obscure ways. Hence the adjective ‘sibylline’, which refers to messages that are not immediately intelligible or remain cryptic tout court.

Fast forward to today and the world of copyright and you see that, in Rome, the ancient sybils appear to have been oddly replaced by the Court of Justice of the European Union (CJEU), whose judgments – rather than prophecies – seem to be difficult for national judges to comprehend fully and, therefore, apply correctly.

The thing, however, is that – unlike the famous Cumaean Sybil’s messages – the CJEU judgments are very much in the present and are way more intelligible. As such, they should not face the difficulties and – bluntly put – resistance that, instead, they have been encountering across multiple national courts.

The recent and somewhat ‘twin’ decisions of the Rome Court of First Instance in RTI v Vimeo (decision 5700/2023) and RTI v V Kontakte (decision 14531/2023) are examples of both a misunderstanding and misapplication of CJEU case law, notably the 2021 judgment in YouTube, C-682/18 and C-683/18 [IPKat here].

How we got to YouTube

To understand YouTube, it is necessary to frame it within its broader context, notably the earlier judgment in Ziggo, C-610/15 concerning the infamous Pirate Bay [IPKat here]. There, the CJEU ruled for the first time that a platform operator could be liable – on a primary (direct) basis – for acts of communication/making available to the public within the meaning of Article 3 InfoSoc Directive by facilitating the finding of unlawful content.

Following that ruling, a debate ensued as to whether the holding that internet platform operators may, in certain conditions, be held primarily liable for performing copyright-restricted acts could be extended to less egregious scenarios than piracy-focused ones.

Despite the contrary view of Advocate General (AG) Saugmandsgaard Øe, the Grand Chamber of the CJEU answered ‘yes’ in YouTube, focusing – as the determinative benchmark – on whether the platform operator’s role may be regarded as indispensable and deliberate.

To this end, it is necessary to consider if the platform operator at issue:

• Refrains from implementing appropriate technological measures that can be expected from a diligent operator in the specific circumstances at issue to counter credibly and effectively copyright infringements on its platform;
• Participates in selecting protected content unlawfully communicated to the public;
• Provides tools specifically intended for the unlawful sharing of protected content or knowingly promotes such sharing, which may be attested by the fact that that operator has adopted a financial model that encourages users of its platform unlawfully to communicate protected content to the public via that platform.

The mere circumstance that a platform operator has abstract knowledge that unlawful content may be shared by its users is insufficient, and so is the circumstance that it operates on a for-profit basis. Nevertheless, in either case, the assessment leads to a different outcome if a rightholder has provided a specific notification and the platform operator has refrained from acting expeditiously to remove or disable access to that content.

In sum, YouTube clarifies two key things:

• The first is that the possibility of holding a platform operator directly liable for acts of communication to the public in relation to content uploaded by its users is not limited to egregious scenarios (in this sense, there is material continuity between CJEU case law on Article 3 InfoSoc Directive and Article 17 DSM Directive (adopted two years prior, in 2019, though the latter is not a codification of the former);
• The second is that, in order to establish liability, a multi-factorial assessment needs to be undertaken. Importantly, such an assessment should not be intended as a formal ‘checklist’, but rather as entailing a balancing exercise.

The Rome decisions

Without commenting on the outcome of either case (the platforms at issue – Vimeo and VK – were ultimately held not liable), the Rome decisions are problematic because they misunderstand the CJEU YouTube judgment. This is so for the following reasons.

Misunderstanding of YouTube

First, the Rome Court of First Instance saw an incompatibility between the CJEU holding in YouTube and earlier case law of the Rome courts (as well as the Italian Supreme Court, notably decisions 7708/2019 and 7709/2019 [IPKat here]), which had instead pre-dated YouTube by establishing the primary liability of certain platforms even outside egregious scenarios à la Pirate Bay.

The Rome court misinterpreted the CJEU, claiming that, there, the Grand Chamber would have indicated that activities like filtering, selection, indexing, organization, classification, aggregation, evaluation, use, modification, extraction, or promotion of user-uploaded content could not trigger liability – in contrast with the earlier Italian decisions..

This is incorrect. The CJEU qualified all these activities by (i) distinguishing between lawful and unlawful content (not any content) and ultimately (ii) only endorsing those aimed at countering credibly and effectively the presence of the latter type of content on the platform at hand.

Furthermore, YouTube concerned the primary liability of a platform operator not, as the Rome court appears to think, a situation in which a platform operator might become liable on a secondary basis due to the inapplicability of the hosting safe harbour.

In sum: the Rome court misinterpreted the holding and place of YouTube when it concluded that such a decision has ‘deeply changed the approach to [the treatment of] hosting provider[s] as [previously] emerged within the Rome Court of First Instance […], as upheld by the Court of Appeal of Rome and the
Supreme Court’.

Reduction of YouTube evaluation criteria to formal checklist

Secondly, the Rome court unduly reduced the guidance in YouTube to a formal (and empty) checklist, and even failed to apply it in its completeness.

By reading the judgments, the primary liability of the platform operators in question was ruled out by considering that they had:

• Stated in their terms not to infringe copyright and other third-party rights;
• Automatically – instead of manually – indexed content;
• Required users to create an account;
• Had a notice-and-take down functionality; and
• Most of the revenue generated derived from user subscriptions rather than advertising.

This, as detailed, is not what YouTube mandates.

IPKat-approved checklist ...

What is required, instead, is a weighed considerations of all relevant circumstances, including evaluating the efforts made by the platform operator to counter credibly and effectively the spread of unlawful content, the amount of unlawful content effectively present on the platform vis-à-vis lawful content, and the platform’s overall business model.

Primary (direct) and secondary (accessory/indirect) liability

Third, the Rome court did not correctly separate questions of primary (direct) and secondary (accessory/indirect) liability, with the hosting safe harbour under the Italian equivalent of Article 14 of the Ecommerce Directive (now Article 6 DSA) being only relevant to the latter.

Admittedly, in Italian case law, the divide between primary and secondary liability in relation to online intermediaries has remained ambiguous. Of course, this ambiguity is not exclusive to Italy.

Furthermore, things have been made more complex by the objectively wide reach of CJEU case law and the resulting construction of the right of communication/making available to the public, as well as the persistent uncertainties regarding who and what the safe harbours would be for. In this sense, it is worth recalling that, in his Opinion in YouTube, AG Saugmandsgaard Øe held that the hosting safe harbour would apply ‘horizontally, to all forms of liability.’ Such a position is however incorrect if taken as to encompass own acts of the provider too.

All the above is clear having regard to the wording of the Ecommerce Directive and CJEU case law dating as early as Google France and Google, C-236/08 to C-238/08. In YouTube, the Grand Chamber clarified once and for all that the safe harbours are only available to intermediary service providers. In other words: the safe harbours do not apply irrespective of the type of liability at hand.

If a national court found the platform operators at issue primarily liable for copyright infringement, they would be automatically disqualified from the hosting safe harbour. All this said, if a platform took proactive steps to avoid that infringing activities are performed by users of its service, this should not automatically exclude the application of the immunity.

Such a conclusion is now crystallized in legislation: it is sufficient to look at Article 17 DSM Directive and Article 7 DSA. They clarify that that there is no safe harbour for platform operators holding primary responsibility/liability (Article 17 DSM Directive) and, where an intermediary is eligible for the safe harbours, taking steps to avoid infringements by users should not be a penalizing factor for such safe harbour eligibility (Article 7 DSA).

Yet, the Rome decisions conflate questions of primary and secondary liability and unduly consider that, if the platform is not primarily liable, then it is not secondarily liable either because the safe harbour would apply. This is by no means an automatic progression.

Content of notices and monitoring obligations

Another point regarding these decisions is that the Rome court ‘departed’ from its earlier case law, with regard to the content of notices. In the past, the Rome courts had indicated that a notice does not need necessarily to indicate the URL at which the allegedly infringing content is to be found. With the most recent decisions, the Rome court aligns itself with the practice of other Italian courts (e.g., Milan and Turin), which have mandated a URL-indication requirement.

This, in itself, is not incorrect given the specific circumstances at hand. That said, it is incorrect to think that there is an obligation to indicate the URL in each and every case. In this sense, it is worth recalling that:

• In 2019, the Italian Supreme Court also concluded that a notice must indicate the relevant URL only when this is ‘indispensable’ to identify the infringing content;
• The law does not require necessarily the indication of the URL at which the allegedly infringing content is to be found for a notice to be considered validly submitted. In this sense, the Rome court erred once again when it held that the Grand Chamber in YouTube ‘clearly required’ the indication of the URL at which the allegedly infringing content is found.

While the Ecommerce Directive did not specify what content notices must have, Article 16 DSA now provides a harmonized approach to notice and action mechanisms for providers of hosting services, by requiring notices to be ‘sufficiently precise and adequately substantiated’. This language codifies settled CJEU case law, notably L’Oréal, C-324/09, and provides for an obligation to indicate clearly the exact electronic location of the allegedly infringing content, such as the exact URL or URLs.

... and the only IPKat-approved palm leaves

On a final note, the Rome court deemed the prohibition of general monitoring obligations under Article 15 of the Ecommerce Directive (now Article 8 DSA) as all-encompassing, without also considering the distinction between general (prohibited) and specific (allowed) obligations. In YouTube, the CJEU considered favourably the use of technologies that a diligent economic operator could be expected to adopt in order to counter credibly and effectively copyright infringements on its platform. Such technologies entail monitoring.

Conclusion

The reason why the Cumaean Sybil’s prophecies were so challenging to decipher is because, inspired by the gods, she used to transcribe them in hexameters on palm leaves, which were subsequently mixed up by the winds entering the cave close to Naples she operated from.

Thankfully, this is not how judgments are drafted in Luxembourg. Hence, the difficulty of national courts in applying them appears hard to justify. Hopefully, future cases will be more respectful of the CJEU gods guidance.