Google under EC scrutiny for exploiting content from YouTube publishers and creators in AI services like Gemini

Comissão Europeia announced on Tuesday, December 9, 2025, on Bruxelas, the opening of a formal investigation against Google for possible violations of União Europeia’s competition rules. The focus is on using online content from web publishers and YouTube creators to train artificial intelligence models like Gemini. Autoridades suspect that the company, controlled by Alphabet, imposes unfair terms on publishers and blocks rivals from accessing similar data.
This action comes after complaints from independent media organizations, which allege losses in traffic and revenue. Comissão intends to assess whether Google grants itself privileged access to materials, disrupting the generative AI market. Teresa Ribera, European Commissioner for Concorrência, highlighted that innovations in AI benefit society, but cannot ignore fundamental principles of equity.
Google responded through a spokesperson, stating that the investigation could inhibit progress in an increasingly contested sector. The company emphasizes partnerships with creative industries to adapt to the AI era.
- Main targets of the investigation:
- Use of articles and websites without remuneration or opt-out for publishers.
- Model training with YouTube videos, where creators obligatorily give up rights.
- Barriers to rival developers, preventing use of similar content for AI.

Web content usage practices
Web publishers report that Google incorporates their materials into AI-generated summaries, such as AI Overviews, displayed at the top of search results. Essa feature, launched in May 2025, appears in more than 100 countries and has included ads since then. Sem compensation, publishers face a drop in visits, which affects financial sustainability.
Comissão examines whether such summaries violate standards when prioritizing Google’s own services. Rivais argue that this creates a vicious cycle, where the technology giant strengthens its dominant position without sharing benefits. In July 2025, an alliance of independent publishers filed a formal complaint, speeding up the regulatory process.
Internal data reveals that Google processes billions of pages daily to refine AI algorithms. However, current policies do not provide for automatic payments to original sources, unlike one-off agreements with large media outlets.
This discrepancy raises questions about transparency in training datasets, especially with the entry into force of AI Act in 2026, which requires disclosure of data sources.
Analysis of YouTube Policies
YouTube creators find that by uploading videos, they grant broad license for use in machine learning and AI applications. Essa clause, updated in September 2025, requires the transfer of data for improvements to Google products, including Gemini. Relatórios indicate that videos served as the basis for models like Veo3, with no opt-out options for individuals.
The platform justifies its use as essential to improve user experiences, but critics see it as an anti-competitive barrier. Desenvolvedores of AI rivals, such as those behind open-source tools, do not access the vast collection of 500 hours of uploads per minute, limiting external innovation.
In June 2025, CNBC exposed how educational content and tutorials fuel training without credit or compensation. Comissão is now requesting internal documents to map data flows, potentially imposing fines of up to 10% of Google’s annual global revenue, estimated at billions of euros.
This stage of the investigation could redefine terms of service for video platforms, balancing creators’ rights with technological demands.
Regulatory history with Google
X__NM0____
This new probe marks the second against Google in less than a month, following scrutiny over search spam policies. European Autoridades intensify actions against American big techs, in line with Digital Markets Act, which aims to heal digital markets.
Ribera reinforced that healthy information ecosystems depend on resources to produce quality content. Gatekeepers like Google cannot dictate unilateral conditions, according to an official statement.
Internally, Google adjusted strategies, such as ceasing exclusive agreements with device manufacturers for search and AI apps, opening up space for competitors.
Implications for the AI market
AI rivals, including European startups, celebrate the initiative as a chance to level the playing field. Gemini, integrated into searches and apps like Chrome, processes multimodals such as text, images and video, but depends on massive data that few access.
The investigation may force Google to offer opt-outs and compensation, similar to negotiations with Meta over WhatsApp blocks. Analistas predict that resolutions require sharing of anonymized datasets, fostering competition.
In the short term, publishers are monitoring impacts on advertising revenue, which fell by up to 20% in tests with AI Overviews. Longo term, AI Act will reinforce risk and transparency obligations, affecting global training.
This dynamic reflects transatlantic tensions, with the US criticizing European regulatory rigidity, but Comissão prioritizes protection for local innovators.
Gemini Technical Details
The Gemini model has evolved since initial releases in 2023, incorporating advanced reasoning capabilities tested in 57 academic areas. Versões as well as Gemini 3.0 outperform benchmarks against GPT-4 in 30 out of 32 metrics, according to Google.
Training involves billions of parameters, processed in European data centers for partial GDPR compliance. However, reliance on public content raises ethical debates about intellectual property.
Rival developers seek equitable access, arguing that exclusivity perpetuates monopolies. The probe may require independent audits of data flows, changing big tech practices.
In short, the case tests the limits of innovation versus regulation, with outcomes expected within months.






