20th August 2026: A cultural shift towards safety by design technology accessed by children, new frameworks addressing AI chatbot safety, more consistent parental controls, age-assurance approaches, and child safety certification programs are among the areas for action identified in a new paper exploring how to create a healthy digital childhood.
Growing up in a Digital World has been published by BSI, the global business improvement and standards company, following a roundtable involving policy experts, officials and representatives from industry, advocacy groups and the third sector, exploring how standards can support safety as technology transforms childhood.
The timing is particularly relevant in the U.S., where pressure is growing to build child safety into digital products from the outset. In April, bipartisan Senate legislation was introduced that would require AI chatbot providers to offer family accounts, obtain parental consent and provide protective default settings for teens. Meanwhile, Meta is facing a major trial brought by U.S. states alleging that the company’s social platforms were designed to keep children engaged despite known risks to their wellbeing.
The session covered the mixed messages being sent to children and parents, with educational technology embraced while other tools are restricted. Attendees rejected the idea of a single model digital childhood, but called for a rights-based, child-centered framework that supports different stages of development, with persuasive design, addictive engagement loops and streaks highlighted as problematic. Concerns were raised over "cliff edge" age barriers, with support instead for graduated access models, like those used in advertising.
Those assembled agreed that despite industry concerns that change could have unintended consequences, a cultural shift to prioritize children’s safety is possible. Safety must be built into digital from the outset rather than addressed only after harms emerge. Children should not be prevented from accessing technology, but, as with physical products including toys, tools should be risk-assessed, tested and designed to mitigate foreseeable harms before deployment. Developers, platforms and service providers should be accountable for ensuring products are safe by default.
Attendees also agreed the tech industry urgently needs to take a more active role now, avoiding the delays seen with social media.
Laura Bishop, Digital Sector Lead Artificial Intelligence & Cyber Security, BSI said: “A healthy digital childhood would empower children as well as protect them, giving them the skills, confidence and understanding to navigate digital environments, which they will have to do as adults. An overly protectionist approach reliant on access limits or blocking features risks creating children who are shielded from technology rather than prepared for it. We need a shift towards a culture of safety by design, alleviating pressure on parents or children to overcome highly sophisticated behavioural design, and instead ensuring that child safety is built into digital from the outset rather than addressed only after harms emerge.
We’re entitled to expect digital products to be safe the same way we expect physical ones to be safe. Standards provide a practical route to get there.”
Attendees called for tougher incentives for industry action, recommending that BSI explore a certification or trust-mark model like the internationally used cybersecurity scheme ISO/IEC 27001. This would encourage industry to demonstrate compliance with relevant standards, for example around age assurance, privacy or safety by design.
A similar model could give U.S. policymakers and technology companies a practical way to demonstrate that agreed safety expectations are being met, while providing parents and consumers with greater confidence in the products their children use. Rather than developing new requirements for every emerging technology, standards can establish a common baseline for risk assessment, testing, transparency and safety by design, while evolving with technology.
The session followed BSI research that found half of children already have an AI-enabled toy or learning device outside of school. Yet, as with social media, parents are concerned about safety, security and privacy, with growing threats including AI-generated child abuse imagery, deepfakes, sexual extortion and manipulative chatbots.
The findings highlight a broader challenge for policymakers and technology companies: children are increasingly expected to navigate digital environments designed to maximize engagement, while responsibility for managing the resulting risks often falls on families.
Hannah Swirsky, Head of Policy, at the Internet Watch Foundation said: “AI is becoming more pervasive, even making its way into our children’s toys and games. Yet even as we see the tech develop and improve, safety measures are lagging woefully behind. Children deserve to be able to use the internet and technology in safety, but we have repeatedly seen examples of new tools and platforms which have been released without being rigorously designed with safety built in from the first step. As a result, we have seen children made less safe by AI, both online and offline.
“As we see the amount of AI generated child sexual abuse material surge online, we need to see a revolutionary shift in attitude from tech companies towards prioritising the safety of users, particularly children. Government must make sure there are stringent measures in place to ensure children’s safety improves at the same rate AI tech does.”
BSI is urging governments to draw on an existing internationally agreed standard, BS ISO/IEC 27566-1, Information security, cybersecurity and privacy protection of age assurance systems as they develop approaches to age assurance that protect children without creating unnecessary privacy or security risks.
The framework, published in February and with related guidance due imminently, clarifies how to introduce age checks without this being overly burdensome for people or technology companies, or increasing risks through data harvesting. The standard is already available for use and was initiated and developed with significant input from UK experts.
If adopted by governments, this approach could enable age assurance without requiring social media users to surrender unnecessary personal data, while helping to protect young people from unsuitable content and services and providing greater confidence in the age-assurance system.
