The European Commission has found that Meta infringed on the Data Protection and Digital Markets Act (DMA) by not providing users with a "consent or pay" option for Facebook and Instagram between March and November 2024, as required by law. In response, Meta introduced a new model in November 2024 that uses less personal data to display ads, offering an alternative to free personalized ads for those who do not consent. The European Commission is now assessing the compatibility of this new model with the DMA.
https://natlawreview.com/article/european-commission-fines-apple-and-meta-non-compliance-dmaMeta's AI chatbots have raised concerns over allowing users to engage in sexually explicit role-play scenarios, including with minors, due to lax oversight. The company has partnered with celebrities like Kristen Bell and John Cena for their voices, but these deals did not guarantee the use of their voices for such content.
https://www.techspot.com/news/107697-meta-accused-allowing-ai-bots-engage-sexually-explicit.htmlNscale, a London-based startup less than a year old, is seeking $2.7 billion to build artificial intelligence infrastructure with the help of a partnership with ByteDance Ltd., owner of TikTok. The funds would be used to create data centers globally filled with Nvidia Corp. chips, which would be rented out to companies like ByteDance for AI model training and operation.
https://news.bloomberglaw.com/capital-markets/ai-startup-nscale-chases-bytedance-deal-and-1-8-billion-in-debtPatria highlighted the importance of EU regulations on digital services and markets, specifically mentioning the Digital Services Act (DSA) and Digital Markets Act (DMA). The country aims to adjust these frameworks while emphasizing the need for effective management of platforms, particularly in regards to artificial intelligence (AI), cybersecurity, and digital transformation.
https://en.antaranews.com/news/353173/indonesia-studies-eus-digital-regulations-for-governance-framework?utm_source=antaranews&utm_medium=desktop&utm_campaign=top_newsPublishers are struggling to distinguish between AI chatbots like OpenAI's ChatGPT and search engine bots, as both can crawl and collect information from their content without explicit permission. Google's Bard and Microsoft's Bing have not yet turned off the bot feature that allows them to do this. Publishers can set a "do not crawl" flag, but it may not be effective against large language models, which can bypass this standard. The lack of a unified "do not crawl" technology makes it difficult for publishers to selectively block crawlers, leaving them vulnerable to AI-powered content scraping.
https://digiday.com/media/why-protecting-paywalled-content-from-ai-bots-is-difficult-business/Sentra, a global leader in cloud-native data security for the AI era, has launched its Data Security for AI Agents solution to address emerging challenges with proliferating AI assistants. The solution is designed to empower large enterprises to embrace AI innovation securely and responsibly. With platform support for Agent toolkits including Microsoft Copilot Studio, Amazon Bedrock, and OpenAI ChatGPT Enterprise, Sentra's solution safeguards sensitive information within autonomous AI workflows, giving security teams peace of mind. CEO Yoav Regev stated that organizations must ensure AI agents are used securely to protect the privacy and integrity of underlying data. The new solution provides comprehensive protection for AI copilots and knowledge bases with capabilities such as automatic discovery, intelligent labeling, and real-time monitoring. Sentra's Data Security for AI Agents is purpose-built to secure AI-powered assistants and prevent unauthorized disclosures, allowing large enterprises to confidently embrace AI innovation while protecting valuable data assets.
https://www.prnewswire.com/news-releases/sentra-introduces-data-security-for-ai-agents-solution-to-secure-the-next-generation-of-business-automation-302438885.htmlTencent Holdings and Alibaba Group Holding have purchased a substantial number of graphics processing units (GPUs) from ByteDance, which previously stockpiled about $13.7 billion worth of Nvidia's H20 chips. The GPUs were primarily used to support the development of artificial intelligence (AI), including ByteDance's ChatGPT-like application Yuanbao.
https://www.scmp.com/tech/big-tech/article/3308254/tencent-alibaba-buy-nvidia-gpus-bytedance-stockpile-report-says?module=top_story&pgtype=sectionThe US government has dismantled key guardrails against disinformation under President Donald Trump, with slashed funding for research and the closure of a key agency combating foreign influence operations. The National Science Foundation cancelled hundreds of grants focused on diversity, equity, and inclusion, as well as misinformation and disinformation, citing cost savings of $233 million. This move has left researchers concerned about the spread of false information, particularly in health and tech sectors. The State Department's Counter Foreign Information Manipulation and Interference hub was also shut down by Secretary Marco Rubio, which had tracked and countered disinformation from foreign actors. Experts warn that this could grant US adversaries more freedom to sow disinformation, as social media platforms scale back content moderation and Meta suspends third-party fact-checking in the US.
https://www.thestar.com.my/tech/tech-news/2025/04/28/us-anti-disinformation-guardrails-fall-in-trump039s-first-100-daysA new licence aims to provide legal certainty and protection for creators by simplifying access to data for training GenAI models and ensuring fair compensation for their works. This move benefits smaller creators who previously lacked bargaining power, while also potentially reducing the number of copyright litigation cases worldwide. The availability of such licences in the UK and internationally could alter the risk profile for AI system transactions, addressing a long-standing concern about third-party copyright litigation risk.
https://natlawreview.com/article/could-be-ai-nswer-collective-copyright-licence-generative-ai-training