Here’s a preview of what you’ll find in this month's SafetySpace edition:
🌐 Regulatory Brief: The 🇪🇺 European Commission preliminarily finds Meta in breach of the DSA for not preventing minors under 13 for accessing their services, and has set its eyes on ChatGPT as a VLOSE. 🇬🇧 Ofcom opened three new investigations. 🇦🇺 Australia's eSafety Commissioner geared up to sue platforms over weak child-safety measures. And there’s more to share.
💻 Inside the Tech: X integrated Grok into photo edits after its image generation function got highly scrutinized. Meta is allowing parents to see their children’s conversation records with Meta AI. PlayStation rolled out new age checks to limit unverified users from its communication features. And Roblox is launching new age-based account tiers.
🚦 In case you missed it: A new UNICEF report on how governments and platforms can act in the interests of children in the digital space, a new comparative study on online safety regulations across 19 jurisdictions, and research on risk-scored moderation and AI-powered child protection. We also published an article breaking down your new CSEA reporting duties, apractical guide that helps you avoid common pitfalls in implementing AI moderation tools, and an article unpacking the new T&S and compliance challenges of chatbots. And there’s more not to be missed.
🇪🇺 The European Commission, the Parliament, and member states continue advancing on age assurance, child safety, and further platform obligations this month.
📘 The European Commission preliminarily finds Meta’s measures to prevent minors under 13 from accessing Instagram and Facebook inadequate under the DSA, and the risk assessment methodology behind them insufficient.
📋 The Commission is also likely to expand on the number of VLOPs in the near term, as Roblox and Reddit have passed the 45 million active monthly users threshold in the EU. In addition OpenAI's ChatGPT is likely to be designated a Very Large Online Search Engine (VLOSE) – a move that would impose stricter transparency and risk management obligations on the chatbot.
🔴 But the DSA framework itself is being challenged:Google is seeking to rebut the EC’s refusal to de-designate Google Shopping as a VLOP, raising questions about how “average monthly active recipients” should be calculated.
📵 EU member states are signaling openness to align their strategies on social media restrictions for minors. Joined by member states and EC President Ursula von der Leyen, French President Emmanel Macron has actively pushed for a unified approach to strengthen minors’ protection and hold major online platforms more accountable.
🔍 The European Parliament has voted against prolonging a temporary derogation to the ePrivacy Directive that permits online platforms to scan for CSAM on their services, which has expired on 3 April. Major tech firms including Google, Meta, Snap, and Microsoft warned that the lapse has left them without a clear legal basis to run automated CSAM detection on their services.
🇬🇧 Ofcom opened three new OSA investigations, a new statutory CSEA reporting duty went live, and the Commons advanced proposals for the under-16 social media access restriction.
🔎 Ofcom opened three investigations into Telegram and teen chat sites, probing whether the providers have met OSA duties on illegal content and child safety, particularly CSAM/CSEA risks. Prime Minister Keir Starmer also urged executives from Meta, Google, TikTok, Snap, and X to strengthen child-safety measures, signaling continued pressure on platforms to demonstrate how they protect minors.
🚨 The UK has launched the CSEA Industry Reporting Portal (CSEA-IRP) that requires in-scope platforms to report Child Sexual Exploitation & Abuse (CSEA) content in line with the Online Safety Regulations 2026. The new reporting duty 1) applies to all platforms regardless of service size or CSEA risk level, 2) concerns all CSEA content linked to the UK, and 3) may be satisfied through reporting to a foreign agency, such as NCMEC, but only if all UK-linked CSEA content is reported at the foreign agency. Check out our latest article to see how this new duty may impact your services.
🏛️ The Commons rejected a Lords amendment that would have blocked the under-16 social media access restriction. The government agreed on alternative amendments that would enable the Secretary of State to extend the restriction to social media platforms. In-scope services could be obliged to implement stronger age verification mechanisms and restrict certain features (e.g. infinite scroll, AI chatbots) on their platforms.
🌏 APAC and US regulators continued to test how platforms operationalize online safety and content moderation duties.
🇦🇺 Australia's eSafety Commissioner signalled possible legal action against Meta, TikTok, Snapchat, and YouTube over potential failures to block users under 16 - a significant escalation that could move matters to court and statutory fines.
🇸🇬 Singapore's Infocomm Media Development Authority (IMDA) brought its Code of Practice for App Distribution Services into force on 1 April 2026, requiring designated app stores, including Apple App Store, Google Play, Huawei AppGallery, to deploy age assurance so under-18 users cannot download age-inappropriate apps. IMDA has signalled it will extend age assurance to designated social media services next.
🇺🇸 In the US, a bipartisan team of senators met with child-safety advocates to advance three federal proposals, including a proposal allowing victims of CSAM to sue online platforms and one that demands stronger online safety guardrails for minors.
🤖 X rolled out AI-powered automatic post translation and a new photo editor in its post composerthat can modify images via text prompts, both powered by xAI's Grok model. Although X restricted the image generation feature to paying users earlier this year due to regulators' pressure, this new AI powered image editing feature is available to all users - currently on iOS and with Android to follow soon.
🔷 Meta is allowing parents to see what topics their children have asked Meta AI on Facebook, Messenger, or Instagram. It’s adding a new “Insights” tab within the supervision hub showing the topics their children have been discussing with the AI chatbot.
🎮 PlayStation will require age verification to use messages and voice chat later this year. Users who skip verification would keep access to games, trophies, and the store but lose communication features. The change follows similar age-assurance moves by Discord and Roblox.
🧒 Roblox is launching new age-based account tiers - "Kids" (5–9) and "Select" (9–15) - thatrestrict games and chat by age, which will roll out globally in June. The tiers build on mandatory age checks introduced in January and arrive amid lawsuits from US state attorneys general over child safety.
In-scope user-to-user services in the UK are facing new duties to report detected Child Sexual Exploitation and Abuse (CSEA) content. But what actually triggers a report? When is NCMEC reporting enough? And where do T&S workflows need to change? This article breaks down what teams should assess before enforcement begins.
Integrating third-party AI moderation tools into T&S workflows requires clear evaluation: what data trained the model, what pre-integration testing is needed, and how AI and human review should coordinate. This article dives into key questions to ask, realistic performance expectations today, and what it takes to close the automation gap. Read the full article here.
Online safety frameworks are extending to cover AI chatbots. Within weeks of 2026, Ofcom and the European Commission opened investigations into X's Grok, and Ofcom separately probed AI companion Joi AI. This article unpacks how chatbots complicate risk mitigation for the T&S and compliance teams and what platforms' governance strategy should look like going forward. Check out the full article here.
🇺🇳 The UNICEF published a report outlining how governments and platforms should actin the best interest of children in the digital environment. The report sets a list of priorities grounded in children’s interests, alongside actionable recommendations for governments and technology companies.
🚸 Internet Watch Foundation (IWF) found 15,031 commercial child sexual abuse sites in 2025, a 114% increase from 2024 (7,028). The organization warned that criminals are taking advantage of systemic failures in online security infrastructure.
🛟 The Tech for Good Institute discusses how Trust & Safety fills the critical gap in the online harm response ecosystem. It positions T&S as the middle ground between slow legal enforcement and public education, giving platforms speed and proactive harm prevention, and calls for a multidisciplinary field with shared skills, ethics, and standards.
🟪 NYU Stern Center for Business and Human Rights published a new report on global online safety regulations. This study analyzes 26 laws across 19 jurisdictions globally. It offers a structured framework for understanding and comparing these regulatory approaches. The report also proposes key considerations to navigate future regulatory development consistent with international human rights standards.
“FlexGuard: Continuous Risk Scoring for Strictness-Adaptive LLM Content Moderation” reframes LLM content moderation from fixed binary classification into a calibrated risk score configurable via thresholds. This is a worthwhile read for online platforms reassessing whether their moderation stack can keep pace with shifting enforcement demands.
“AI Powered Framework for Child Protection and Digital Safety” proposes an AI-powered framework combining NLP, computer vision, and behavioral analytics to proactively detect cyberbullying, grooming, and explicit content targeting minors across digital platforms. For platforms reassessing their approaches to protecting children online, this piece provides a scalable, adaptive, and ethical solution to ensure children’s safety in the digital world.
.png?upscale=true&width=1120&upscale=true&name=Newsletter%20visuals%20(1).png)
