Canada proposes wider safety duties for social media and AI chatbots
Canada’s proposed Safe Social Media Act would require covered social-media and AI-chatbot services to reduce systemic risks, implement child-protective design, use privacy-protective age assurance and publish safety plans for independent oversight.
What it could improve
Bill C-34 would shift responsibility toward the services that design and operate digital systems. Its proposed child-protection duty includes safer design and an under-16 threshold for certain social-media accounts, while a Digital Safety Commission would oversee safety plans, guidance and enforcement. The proposal also addresses chatbot risks, including harmful responses and crisis situations.
What remains unresolved
Broad service duties are promising, but the proposed regime does not automatically deliver consistent child protection throughout the wider Internet.
- The bill leaves some detail to future regulations, including how services will meet design and assurance duties in practice.
- Privacy-protective age checks still need measurable limits on collection, reuse, retention and disclosure of children’s data.
- Private messaging is outside the proposed regime, and children may encounter risks in other services or while signed out.
- A minimum-age rule can be bypassed through inaccurate or shared accounts unless real-world effectiveness is assessed.
- Published safety plans need independent scrutiny and evidence of improved outcomes, not merely descriptions of internal processes.
The Child-Friendly Internet perspective
Canada’s proposal combines safer design, AI accountability, oversight and explicit privacy limits. A Child-Friendly Internet approach could complement those duties with interoperable, data-minimising child assurance, allowing participating services to respond appropriately without each independently collecting the child’s identity.
- Treat age assurance as a limited signal for an age-appropriate response, not permission to build a persistent profile.
- Align oversight with measurable safety, wellbeing and participation outcomes.
- Consider how protection continues when children move between social media, games, AI tools and other services.
The central question
Can Canada’s service-level duties become coherent protection across digital environments without repeated identification of children?