Generative AI challenges the right to be forgotten. Large language models don’t store data as static records, they memorize patterns. Removing personal data isn’t as simple as deleting a row in a spreadsheet. Enter machine unlearning: emerging methods let AI models “forget” specific data without full retraining. Techniques like Example-Tied Dropout (ETD) and Redirection for Erasing Memory (REM) are enabling certified, auditable erasure while respecting computational limits. T

The European Union has officially adopted a new regulation under the Omnibus VII package to simplify and streamline the implementation of the AI Act. The new rules postpone the application of requirements for high-risk AI systems to 2027 and 2028, providing greater legal certainty and reducing compliance burdens for businesses. At the same time, the regulation strengthens safeguards by banning AI systems used to create non-consensual sexual deepfakes and AI-generated child se

As AI transcription tools like Otter.ai, Fireflies, and Microsoft Copilot become ubiquitous in meetings, organizations face a new set of privacy and compliance challenges. Employees may unknowingly record sensitive conversations, HR discussions, performance reviews, or medical accommodations, without proper consent or data safeguards. Organisations should update their AI policies to address recordings and transcripts, specify which meetings should not be recorded, review data

The UK Government has announced a series of measures aimed at enhancing the digital safety of children and teenagers, including a ban on social media access for users under 16, restrictions on certain AI chatbots, and the introduction of a “digital curfew” from 8:30 PM. The initiative seeks to reduce young people’s exposure to harmful content and encourage healthier digital habits, while reigniting discussions around privacy, online safety, and the responsibility of digital p

The Court of Justice of the European Union (CJEU) ruled that pseudonymized data does not always qualify as personal data under the GDPR. The decision emphasizes that, depending on the context, pseudonymization can prevent third parties from identifying data subjects, meaning GDPR rules may not apply in such cases. However, data controllers remain responsible for informing data subjects about processing and any data sharing, even if re-identification by recipients isn’t possib

The European Parliament has switched off built-in AI features on the work devices of MEPs and staff due to cybersecurity and data protection risks. Some AI tools were sending data to cloud services, raising concerns about potential data exposure. While daily apps like email and calendars remain unaffected, the move highlights the growing need for caution when using AI tools, especially in sensitive environments. The Parliament advises staff and lawmakers to apply similar priv