RTP and Observador report that the social network X has implemented measures to stop its AI tool Grok from targeting or producing altered images of real people (including so‑called ‘undressing’ outputs). The change follows criticism and regulatory pressure in several countries and is presented as a platform‑level content‑safety update. Social‑media users and journalists in Portugal should check platform settings and be aware that image‑manipulation features may be restricted or moderated more aggressively going forward.