Impact of regulation of children’s digital lives – Phase II
Following on from Phase 1 of the Impact of regulation of children’s digital lives report, Phase 2 covers 2024–2026, analysing child safety and privacy changes across 70 platforms (including social media, AI chatbots, and gaming) against the backdrop of the UK Online Safety Act, the Age-Appropriate Design Code, and the EU Digital Services Act.
Executive summary
2026 is a pivotal year for regulating children’s safety and privacy. Significant questions are being asked about the effectiveness of existing frameworks, amid debates over new legislation to delay children’s access to digital platforms, and evidence from US litigation alleging that platforms prioritise engagement over safety.
This Phase 2 report examines how regulation has shaped the design and governance of online services likely to be accessed by children between 2024 and 2026. It also assesses progress since the first report, published in May 2024. The initial phase of research analysed the impact of regulation from 2017 to 2024, reviewing design changes across 50 online platforms, with a particular focus on Meta, Google, TikTok and Snapchat. The findings indicated that regulation – especially the UK Age-Appropriate Design Code (AADC) – had driven a number of significant design changes aimed at protecting children, while highlighting the need for further research to evaluate their effectiveness. The study broke new ground as one of the first comprehensive analyses of this issue.
The research focuses on three regulatory regimes: the UK Online Safety Act (OSA), UK Age-Appropriate Design Code (AADC) and the EU Digital Services Act (DSA). These frameworks were selected because they have the most comprehensive scope and are now fully implemented in relation to specific obligations for children. The findings are also intended to inform policymakers in other jurisdictions considering the introduction or reform of similar regulations.
For Phase 2, evidence of change across 70 platforms was examined, spanning social media, AI chatbots and gaming services likely to be accessed by children in the UK and EU. A total of 108 changes recorded between 2024 and 2026 were categorised and analysed. Evidence was gathered through direct requests to platforms for information on safety and privacy measures, alongside a review of public sources, such as press releases, blogs and policy updates. Only five platforms responded, with several large platforms citing a lack of resources. A central element of the Impact of regulation of children’s digital lives: Phase II – 2026 8 analysis focuses on Meta, Google, TikTok and Snapchat –platforms widely used by children – enabling comparison with Phase 1 findings.
Evidence from other research considering the effectiveness of the changes, and the persistence of risks to children, was also considered, along with an analysis of the actions of regulators.
The new measures were categorised into four types: by default, tools, information and support. ‘By default’ safety measures are a critical component of safeguarding children online, as they embed protections into the user experience from the outset without requiring action by children or parents – for example, autoplay being switched off by default. In contrast, tools are user-activated safety features, such as parental controls. Information-based measures provide children and parents with guidance and prompts to support safer use, while support measures focus on mechanisms such as reporting harmful content or behaviour.
Overall, this report illustrates where regulation is continuing to have an impact, including in the use of ‘by default’ measures, particularly enabled by age assurance. Problematically, it reveals that, while significant risks persist online, some major platforms have shifted from providing “by default” protections to providing end-user tools, which, in turn, shifts responsibility for online safety to the child. There is also significant evidence that indicates that risks are still persisting on platforms despite the measures in place. Current regulations retain significant potential to make a major improvement in children’s lives, provided the laws, codes and guidance are further clarified, and backed by strong enforcement. This must also be backed by transparent and effective regime of independent testing and certification. In 2026, we have stronger indications that this enforcement is now finally emerging. Last, the report identifies the new and emerging risks from AI to children and what steps have been taken to address them.
