The comment period for the European Commission’s review of EU copyright rules, including the 2019 Directive on Copyright in the Digital Single Market provided an opportunity to give input regarding the effectiveness of the EU’s copyright framework and the impact of generative AI on licensing, enforcement, and creators’ rights. Comments were closed on June 25.
The responses were diverse. Copyright advocates serving the media and entertainment industry lobbied to codify an obligation for transport providers to respond to of instances of piracy, by blocking infringing sites within 30 minutes (a measure that has had generally positive results but also has its detractors that point to false positives and high costs).
One example, submitted by the Audiovisual Anti-Piracy Alliance, proposed two key regulatory initiatives that should be “urgently taken by European lawmakers that will both provide the desperately needed support for the audiovisual sector against digital live event piracy, and enhance European digital sovereignty and protection for European citizens online.”:
- Provide for a clear legal codification that the requirement on hosting providers to respond expeditiously to take down notices, means in less than 30 minutes1 for live event piracy this will provide the much-needed clarification and even playing field for compliant providers.
- Provide a regulatory means for a competent authority, on application of rights owners, to designate ASNs and IP ranges associated with off-shore non-compliant hosting providers.
Other feedback came from sports leagues, whose agenda is to detect and stop incidents of piracy as quickly as possible, as live sports matches are valuable and short enough to evade detection.
beIN Media France said that “current EU legislation remains insufficient and ill-adapted to address the real-time nature and scale of live content piracy,” and calling for a four-part program:
- Effective Notice-and-Takedown for Live Content: The value of sports rights lies in live transmission, requiring immediate enforcement.
- Know Your Business Customer Obligations: A key enabler of piracy is the anonymity with which operators can access hosting and network resources.
- Duty Not to Facilitate Non-Compliant Offshore Providers: Pirate services rely on a broad ecosystem of intermediaries, many operating through EU infrastructure, and
- Harmonised Dynamic Blocking of Live Piracy: Dynamic blocking is a critical complement to takedown mechanisms, especially against illegal IPTV services, yet its implementation remains fragmented across Member States”
ISPs have their say
Cloudflare’s input was representative of communiations providers, saying that the Commission should “focus on enhancing enforcement of existing legal frameworks rather than issuing new legislation or broader blocking mandates; and advocates for fostering greater cross-border coordination and voluntary industry collaboration.” Reasonable enough.
But the bone of contention lies in the details. “Any enforcement measures should remain proportionate, transparent, and targeted at the correct hosting layer of the Internet stack.” Cloudflare in the past has raised the issue of unintended consequences, including suppressing openness on the Internet, and the cost to implement blocking countermeasures.
Google expanded on the notion of unintended consequences, saying that:
“Italy’s “Piracy Shield” has blocked Cloudflare IPs hosting over 42 million domains, a Google Drive subdomain and legitimate VPN and CDN services. It caused widespread service outages and the exit of VPN providers from the market. Recently, French court orders have been extended to DNS resolvers and VPNs, causing CISCO to withdraw OpenDNS from the French market. A Spanish court-ordered IP block caused third-party platforms, cloud services, APIs, SME sites to be blocked, pursuant to demands from a sports league.”
What about AI?
Artificial intelligence advocates offered a different perspective. Anthropic, for example, contended that
“Copyright restricts certain uses that communicate the protectable, expressive aspects of a work, but it allows learning from works to identify uncopyrightable elements, such as facts and ideas.
“A text and data mining exception (to copyright restrictions) is aligned with copyright’s traditional purpose and function. … Any rights-reservation system embedded in copyright law must be machine-readable, interoperable with existing tools like robots.txt, and feasible for developers to implement at scale. Training-data transparency requirements should … respect the highly confidential nature of dataset composition, and should be interoperable with existing international frameworks” said Anthropic.
Like Anthropic, Microsoft lobbied for the EC to maintain exceptions to copyright law for Text and Data Mining (TDM) Exception under Articles 3 and 4 of the Copyright in the Digital Single Market Directive (CDSMD) (as well as Section 29A of the UK’s Copyright, Designs and Patents Act). Meta made a similar request: “At a minimum, the Article 4 TDM exception must be safeguarded.”
Cloudflare made an interesting observation around data retrieval by AI platforms:
“According to Cloudflare data, 89.3% of AI crawler requests are for model training or mixed purposes that include training 44.6% + 44.7%, while only 8% power a search product that is exclusively referral-driven.” That, according to Cloudflare’s ongoing measurement of AI Bot and Crawler Traffic.
“Only about 8% of AI crawler requests power a search product that sends users back to original content. This one-way extraction of value forces publishers to bear server and bandwidth costs of crawling without receiving traffic or compensation in return. It undermines the sustainability of the creator economy and discourages voluntary licensing and the formation of voluntary content licensing marketplaces,” said Cloudflare.
All of the major AI platform providers supplied input to the EC, including Google, Anthropic, OpenAI and others.
Microsoft’s input was something of a head-scratcher: “It is not yet clear whether the Commission has identified concrete and widespread harm associated with the use of copyrighted content for AI, or whether concerns remain uneven across sectors and use cases, including where many data- driven applications may have little to no relevance to the interests of rightsholders.”
In reality, the EU (EC, EUIPO, and others) have published extensively about the use-cases and harms stemming from copyright infringement and piracy.
Why it matters
Conclusions and decisions reached by the European Commission and European lawmakers may change the contours of piracy / anti-piracy regulation and inform the functionality and design, both of technical solutions and of anti-piracy best-practices
Further reading
Targeted initiative for a better copyright environment for European creativity and innovation. Landing page with links to 432 feedback submissions. Accessed July 7, 2026. European Commission
Why the EU TDM Exceptions (Art 3 & 4) can be used for AI – including generative AI. Article. June 3, 2025. by Benjamin White. Knowledge Rights 21
CCIA Europe: As FIFA competition begins, concerns over fighting piracy without breaking the Internet. Article. June 12, 2026. by Steven Hawley. Piracy Monitor
