AI labelling under Article 50 of the EU AI Act: What applies from August 2026
The transparency obligations under Article 50 of the EU AI Act apply from 2 August 2026. However, the Regulation does not establish a general labelling obligation for all AI-assisted content.
AI-driven trademark abuse: Why “register first, use later” is becoming riskier
AI tools make it easier to identify unprotected brand names. For businesses, this can quickly become costly on platforms such as Amazon.
LG Hamburg: AI-generated X post remains attributable to the account operator
The Regional Court of Hamburg held that a continuing defamatory false statement on X remains unlawful under the law of statements even if the post was generated by AI. The account operator can be held responsible for the published content.
Cologne Higher Regional Court: Meta may provisionally use public Facebook and Instagram data for AI training
The Cologne Higher Regional Court rejected an interim injunction against Meta’s announced use of publicly shared Facebook and Instagram data for AI training. In its summary assessment, the court considered the processing likely lawful, in particular on the basis of Article 6(1)(f) GDPR.
Higher Regional Court of Hamburg: Downloading a photo for an AI training dataset may be permissible under TDM and the research exception
The Higher Regional Court of Hamburg held that downloading an online-accessible photo in order to compare image and description for an AI training dataset may fall within Section 44b German Copyright Act. The court additionally found that Section 60d German Copyright Act could apply because both the dataset creation and its later AI use qualified as scientific research.
Memorisation of AI training data infringes copyright (GEMA vs. OpenAI)
The Munich Regional Court (LG München I) rules in favor of GEMA: The incorporation of copyrighted training data into AI language models constitutes a copyright infringement – injunction, disclosure, and damages sought against OpenAI for training GPT with song lyrics.