A highly notable finding buried in Retraction Watch's weekend reads: hidden prompts (prompt injection attacks) embedded in conference paper submissions were shown to 'snare' AI peer review tools into generating favorable or manipulated reviews. This is a critical integrity story — as AI review tools proliferate, adversarial manipulation of those tools becomes a new attack surface for the scholarly record. Directly relevant to p2preview's core AI + publishing integrity coverage.

Source: Retraction Watch