Like the AI winters of the 1970s and 80s, which were followed by periods of expansion and growth and by exciting developments, new questions, and challenges, we may be experiencing something similar today. Not just with AI, but more broadly in the landscape of online safety. Technological advances, evolving social dynamics, and new ways of interacting online are expanding at an unprecedented pace, opening both opportunities and questions that require a different way of seeing things.

​

Guardianship is not limited to individual actors but produced through a combination of systems, relationships, and institutions. Effective guardianship online emerges from the interaction between algorithmic tools, platform features, social dynamics, and governance structures, challenging the idea that safety can be delivered through human moderation alone.

​

Building on this work, I advocate for a guardianship-based approach to trust and safety that draws on crime science principles, including Routine Activity Approach and Situational Crime Prevention. This approach treats platforms as environments rather than neutral tools, and applies systems thinking across product design, user experience, threat prevention, escalation, and policy. It offers a way to embed guardianship structurally across trust and safety operations, complementing existing technical and operational measures by shaping how risk is anticipated, deterred, and supported across the platform ecosystem.

​

But what would this look like in practice? That’s the exciting part. It’s a story to be written in new spaces, with new tools and across the evolving ways we connect online. If this is something you're interested in, let's chat!​