In this issue: Indonesia is expanding its social media ban to e-commerce platforms, citing reports of children falling victim to online scams and overconsumption. Texas’ AG is suing Netflix for harvesting data from children and deploying addictive design features. And following the implementation of age-verification checks by the UK’s Online Safety Act, an Internet Matters survey found that 46% of UK children described the age-verification methods as easy to bypass.
Texas’ AG is suing Netflix for harvesting data from children and deploying addictive design features. The suit alleges that Netflix engaged in data harvesting by tracking users’ viewing habits, location, and platform interactions to build consumer profiles. The suit also noted that Netflix employed addictive design features, like autoplay, on kids’ profiles. Filed under Texas’ Deceptive Trade Practices Act, the suit requires the company to purge Texans’ data and remove autoplay from kids’ profiles.
Indonesia is expanding its social media ban to e-commerce platforms, citing reports of children falling victim to online scams and overconsumption. Government officials revealed that children were being exposed to age-inappropriate products on platforms and making transactions without parental supervision. “When we talk about PP Tunas, the object of the regulation is all electronic system providers, including marketplaces,” the government’s secretary of the directorate general at the ministry’s digital space supervision unit stated. Indonesia’s under-16 ban, implemented on March 28, 2026, requires companies to implement robust age verification mechanisms, with non-compliance involving fines or blocking of platforms.
The European Commission signaled plans to crack down on “addictive and harmful design practices” on social media platforms, like TikTok, Meta, and X. Citing concerns over depression, self-harm, and suicide linked to social media use, the Commission’s forthcoming Digital Fairness Act (DFA) would ban addictive features and misleading influencer marketing. The Commission is also potentially considering a social media ban for minors. Expected to be proposed by the end of the year, the DFA is intended to strengthen the EU’s existing Digital Services Act framework.
(Exploring features or policies that support kids and teens? Reach out here to schedule a consultation.)
Singapore’s Minister for Digital Development and Information, Josephine Teo, released a Straits Times op-ed arguing that a blanket ban may not be the answer to protecting children. “Bans alone may not bring about real changes in how children interact with social media or develop healthier habits online,” Teo stated. Instead, she expressed the government’s decision to focus on harmful platform features and differentiated offerings for young users, likening them to “training wheels” before children access a platform’s full range of features. In parallel, Singaporean Parliament member Rahayu Mahzam similarly echoed that the government wants “more demanding, more rigorous” regulations than blanket bans.
Following the implementation of age-verification checks by the UK’s Online Safety Act, an Internet Matters survey found that 46% of UK children described the age-verification methods as easy to bypass. Surveying roughly 1,000 children and parents across Great Britain, children reported using VPNs, submitting another person’s face, and using fake birthdates – even drawing fake facial hair to bypass automated systems. Approximately 30% of participating parents also admitted to allowing their children to bypass the age checks. These findings come amid growing frustration among online safety advocates over the UK government’s perceived lack of urgency for child protection reforms; Former Safeguarding Minister Jess Phillips resigned from the government, citing frustration over stalled proposals aimed at preventing children from sending and receiving nude images on their phones.
New Mexico’s Department of Justice is in the remedies phase of its landmark trial against Meta, centered on determining whether Meta’s products constitute a “public nuisance.” The state requested sweeping sanctions against Meta, including a $3.7 billion abatement plan to fund mental health services, law enforcement, and educational programs addressing harms to young users. Prosecutors are also seeking major changes to the company’s systems, including universal age verification, de-encryption of minors’ messages, parental controls, and an independent child safety monitor. Meta has pushed back against the proposal, arguing that the state’s plan would improperly require the company to finance mental health care for all teenagers in New Mexico, not solely those allegedly harmed by its platforms. The trial is expected to last three weeks, following the historic verdict holding Meta liable for endangering children.
Pennsylvania has filed a lawsuit against Character Technologies Inc., the parent company of Character.AI, alleging its chatbots unlawfully present themselves as licensed medical professionals. Described as a “first of its kind enforcement action,” the complaint asks the Commonwealth Court to halt the systems from engaging in “the unlawful practice of medicine” and misleading users into believing they are receiving legitimate medical advice. The case underscores growing concerns about young users relying on chatbots for mental health support instead of clinical services. In response, Character.AI stated that it prioritizes user well-being, noting that the platform includes disclaimers that characters are fictional and should not be treated as real professionals.
OpenAI has introduced a Trusted Contact feature for adults, building upon the company’s parental controls notifications for teens. Trusted Contact allows adults to nominate an individual to be notified when the automated system detects discussions of self-harm that indicate a serious safety concern. While the feature was previously exclusive to teens, the company is “extending [their] safety alert options so anyone over 18 can choose to add someone they trust as their Trusted Contact,” suggesting that features built for teens can also be helpful to the broader population. The company has also developed safety summaries: tightly scoped notes to provide safety-relevant context to models in high-risk situations.
EdTech platform Canvas temporarily shut down after a hacking group claimed to have attacked its parent company, Infrastructure, breaching the data of 275 million students and faculty members across 9,000 schools. Compromised information included personal identifying information, like students’ names, email addresses, ID numbers, and Canvas messages. While Infrastructure was eventually able to restore platform access, the company later informed users that it had paid a ransom to the hacking group to resolve immediate issues. The hack also prompted a lawsuit alleging negligence, breach of implied contract, breach of confidence, and unjust enrichment, with the plaintiff seeking more than $5 million in damages tied to the breach. The Canvas hack follows a string of cybersecurity breaches affecting major EdTech platforms, including the 2024 PowerSchool breach, and arrives amid growing scrutiny over schools’ reliance on technology.
Instagram and Facebook will be using AI to detect underage users, analyzing profiles for contextual clues and engaging in visual analysis – like bone structure – to identify minor accounts. If the company determines an account to be underage, the account will be deactivated and require the account holder to provide proof of age to prevent deletion. The company is also expanding its technology that proactively places suspected underage users into Teen Accounts across Brazil and 27 EU countries. In addition, the company plans to send notifications to parents in the US with information confirming their teens’ ages on Meta’s apps.
The Oversight Board is conducting new research on governing AI chatbots for 13 to 17-year-olds. Given that chatbots are becoming a primary interface for children to access information and express creativity, the Board is requesting public comments on the risks and opportunities of AI chatbots for a forthcoming white paper.
A study by the National Bureau of Economic Research found that cellphone bans in schools have little academic benefit. In a working paper, researchers found that the average effect of the ban on test scores is “consistently close to zero,” with little effect on attendance, in-class engagement, and perceptions of online bullying. Additionally, while teachers reported reduced phone use in class, researchers also observed an uptick in disciplinary incidents and a dip in student well-being in the first year of the ban’s implementation – likely due to student frustration with the new restrictions. “A lot of times, easy solutions seem like they might work really well, [but] it’s hard to move outcomes like test scores,” said E. Jason Baron, one of the paper’s authors.
Boston Children’s Digital Wellness Lab published a survey finding that young users, particularly in Brazil, leverage AI for creative work, personal advice, and social connection (in addition to learning and research as the dominant use cases). A significant portion of teens reported using AI for digital art, creative writing, and graphic design, although there was less consensus on whether AI enhanced or diminished their curiosity. Notably, roughly one-third of teens reported frequently asking chatbots advice on topics like college planning, health, and lifestyle. Among those who acted on AI advice, approximately 80% of US and Brazilian teens claimed that it improved their situation.
More than 70 million warning messages were sent to people attempting to access child sexual abuse material (CSAM) over the past two years as part of Project Intercept, a partnership between child protection charity Lucy Faithfull Foundation and tech firms like Google, TikTok, and Meta. Out of those messages, roughly 700,000 users went on to access Stop It Now resources, which offer self-help tools to support behavioral change. While experts claim the figure is alarmingly low, the Foundation claims that upstream interventions – like warning messages – could help disrupt harmful behavior earlier on to tackle the spread of harmful content.
Silicon Valley-based venture capital firm a16z released a blog post outlining principles to help craft a liability regime for AI – a helpful framework for thinking about accountability across the ecosystem. Of note, they argue that liability should generally rest with users who misuse AI systems, rather than with general-purpose tool developers. Policymakers should also adopt a rebuttable presumption of user liability, alongside narrowly tailored duties of care that avoid overburdening startups. Courts are encouraged to rely on traditional product liability principles, such as risk-utility balancing and the availability of safer alternative designs; however, specialized liability regimes may be necessary in high-risk contexts like mental health and child safety. They further recommend a federal-state approach: Congress should establish a national liability framework, while states retain authority to address harms within their jurisdictions. Finally, policymakers should offer regulatory sandboxes and policy experimentation to allow rules to be refined based on real-world testing.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.