Artificial-intelligence video is moving from one-shot generation toward something much closer to an editing session. On November 5, Google made Gemini Omni 1.2 Flash generally available through its API, business products and several creative platforms. The model can generate clips up to fifteen seconds, hold visual references more consistently and accept conversational revisions to an existing clip.
For artists, the consequential change is not simply five additional seconds. It is the ability to keep a generated sequence alive through successive decisions: change the weather, replace a background, adjust a camera angle or swap an object without rebuilding the entire shot. That turns prompting from a request for a finished result into an ongoing directorial process.
Reference images become production inputs
Google says the model can combine character, product and style images with short motion clips and audio cues. It is designed to keep faces, garments, typography and colors stable across a sequence. In practical terms, an illustrator could animate a character sheet, a sculptor could test a moving presentation of an object, and a filmmaker could develop a storyboard into a rough scene while preserving more continuity between shots.
That same control raises a rights question. A reference image is not just inspiration once it is uploaded to guide identity, dress, style or movement. The artist or studio needs authority to use every face, garment design, artwork, photograph and sound supplied to the system. A model’s ability to follow a reference accurately does not create permission to use it.
Conversational editing changes the evidence of authorship
A finished AI-assisted clip may now emerge through many small revisions rather than one prompt. That can strengthen human direction, but only if the process remains visible. Artists should preserve the initial brief, source files, reference permissions, significant prompts and exported versions. A concise decision log can show which parts were generated, which were selected or rejected, and where human editing changed the result.
This record matters for more than credit. It can help resolve client questions, reproduce a successful sequence, verify that a licensed reference stayed within its approved use and explain the creative contribution if authorship is later disputed.
Provenance is present, but it is not the whole answer
Google says outputs carry an imperceptible SynthID watermark. The model card also notes that the system can alter a person’s speech, and that Google is currently restricting that capability while it studies the risks. Those safeguards are meaningful, yet neither a watermark nor a production filter replaces consent, disclosure or contractual clarity.
Artists should ask where the watermark survives: export, compression, color grading, reframing and re-editing may change a file. They should also decide how an audience will be told that a sequence is synthetic or materially AI-assisted. An invisible signal can help a technical verifier; a visible credit helps a viewer.
What artists can do now
Before using a reference-driven video model, separate materials into three groups: work you own, work you have licensed for this use, and work that must stay outside the system. Obtain explicit permission before generating or changing a recognizable person’s image or voice. Keep original files and non-AI edits alongside generated versions. In a client project, state who approves references, who owns outputs, which tools may be used and whether disclosure is required.
The release shows how quickly generation and editing are merging. The opportunity is greater control over motion, continuity and iteration. The responsibility is equally clear: when a tool can preserve identity more convincingly, artists need a more exact account of whose identity, whose material and whose decisions are inside the frame.
ATIM approaches artificial intelligence with curiosity, independence, and balance. We explore its extraordinary potential to empower artists while examining the questions it raises around authorship, copyright, consent, creative identity, transparency, and human expression.
We are neither advocates for AI nor opponents of it. Our focus is how the technology is used—and what that means for artists.
Above all, our commitment is to artists. We look at every development through the lens of what can support, protect, and empower the creative community—identifying opportunities, examining potential risks, and providing artists with the information they need to make informed decisions about their work, their rights, and their future.