Our automated moderation capabilities are being developed, assessed and introduced progressively. As these systems mature, we evaluate their effectiveness and adapt our processes to address emerging risks.

For security reasons, we do not publish operational details or detection criteria that could be used to evade safeguards.

Automated analysis provides signals that may warrant further attention. Those signals are not definitive judgments and may be affected by the quality or context of the material.

We aim to combine appropriate human review, reporting channels and ongoing improvements rather than relying exclusively on automation.