Ask most technology leaders where their trust and safety risk sits, and they will point to volume. That is the part automation has already solved. The risk has quietly relocated to the decisions your AI systems now make on their own, and to the small share of cases where getting it wrong is a regulatory event rather than a support ticket.
According to PwC’s Trust and Safety Outlook 2026, automated systems now handle the vast majority of routine detections, while the most complex, high‑consequence work is growing.
Regulatory pressure and user expectations are rising in parallel. Under regimes like the EU Digital Services Act and the EU AI Act, robust trust and safety and AI‑safety controls are becoming a regulatory prerequisite for operating large platforms and high‑risk AI systems in Europe. Trust and safety is no longer a back‑office task; it is now a core licence to operate.
The result is a widening gap. There is more volume, more nuance and more risk. There is not enough internal capacity to cover it all. Teams must also build the AI governance, analytics and policy capabilities that platforms now need.
The question for technology leaders is how to create the space to run these operations well. Who handles the high‑volume queues? Who manages 24/7 coverage? Who gives internal teams the breathing space to design the systems that will define the next phase of safe, responsible growth?
This is the gap between where many delivery models currently stop and where trustworthy AI operations need to begin. Covalen's work in scaled trust, safety and AI operations is one reference point for how that gap is being closed in practice.
