<\/p>\n
Faced with demanding exhibition specs and tight deadlines, artists are increasingly using frame\u2011interpolation tools as a last\u2011mile fix \u2014 embracing hybrid workflows that smooth motion while studios, unions and audiences debate the ethics, labour impacts and risks of stylistic dilution.<\/p>\n<\/div>\n
Alice Bloomfield remembers the moment with the kind of wry disbelief that has become common in an industry caught between labour\u2011intensive craft and the rapid promises of generative tools. After three months hand\u2011drawing a stop\u2011motion style music video for screening at Outernet London, Bloomfield told a Nicer Tuesdays audience that the venue requested the project be played back at twice the frame rate. \u201cI\u2019d just spent three months in my room, working, and now you want twice as many frames?\u201d she said, describing the panic of being asked to effectively double her workload at the eleventh hour. The solution, she recalled, was to work with an AI specialist to interpolate a new frame between each existing frame and then do light manual clean\u2011up \u2014 a hybrid fix that preserved the hand\u2011made look while meeting the technical brief. (Sources informing this account emphasise that the interpolation step smoothed motion for the immersive display rather than replacing the original hand work.)<\/p>\n
Bloomfield\u2019s experience is far from an isolated curiosity: generative AI has seeped into visual effects and post\u2011production pipelines across film and advertising. As VFX supervisor Jim Geduldick told The Guardian, \u201cEverybody\u2019s using it. Everybody\u2019s playing with it.\u201d Industry conversations now routinely move between excitement for new workflows and anxiety about what those workflows are displacing, from entry\u2011level jobs to long\u2011established crafts. According to commentators and practitioners, the prevailing pattern is not wholesale replacement so much as practical augmentation \u2014 teams using AI where it saves time, while reserving human judgement for nuance and creative intent.<\/p>\n
Many practitioners frame the best outcomes as deliberately hybrid. Directors and studios cited in the original report say they build bespoke datasets from hand\u2011drawn or filmed assets so that models respond to a project\u2019s particular aesthetic. Studios such as Unveil and animators like Jeremy Higgins have described workflows where human input \u2014 whether the initial drawings, the curation of training material, or final frame\u2011by\u2011frame clean\u2011up \u2014 is the thing that keeps work from feeling hollow. The argument is straightforward: the machine can generate motion or fill gaps, but the human touch supplies the \u201csoul\u201d that audiences read as authenticity.<\/p>\n
Technically, the interpolation Bloomfield used is part of a growing suite of AI\u2011based frame interpolation tools that sit between older optical\u2011flow methods and true learned motion prediction. Technical write\u2011ups explain the basic pipeline: analyse motion across input frames, predict intermediate frames and render them in a way that handles occlusions and complex motion more robustly than many traditional algorithms. Popular models and toolchains \u2014 referenced by engineers and technical blogs \u2014 include approaches like RIFE and DAIN; their practical applications range from slow\u2011motion effects and restoration to smoothing animation for high\u2011frame\u2011rate displays. Developers now talk about integrating these capabilities into streaming players and mobile editors to allow smoother playback and accessibility features in real time.<\/p>\n
The practical impetus for some of these fixes is driven by exhibition hardware. Outernet London\u2019s specifications underline why a 12\u219224fps change matters: its internal LED canvas is a monumental, multi\u2011storey surface with extremely high pixel density and controller infrastructure designed to render hundreds of millions of pixels in real time. On such a surface, judder and motion artefacts are amplified; a higher apparent frame rate and interpolated motion can materially alter how handcrafted work reads to audiences in an immersive setting.<\/p>\n
But the same technical affordances that enable last\u2011minute rescue work also intensify industry\u2011wide concerns. Practitioners warn of two related risks: the normalization of lower\u2011quality, mass\u2011produced content \u2014 often derided as \u201cslop\u201d \u2014 and the erosion of work\u2011opportunities that sustain emerging artists and technicians. Those anxieties echoed loudly during the 2023 actors\u2019 and writers\u2019 industrial disputes, when unions raised protections against the unconsented replication of performers and the use of synthetic replacements for entry\u2011level roles. Union negotiators, industry lawyers and many creatives framed the debate as one about consent, fair pay and the preservation of routes into the industry as automation accelerates.<\/p>\n
Public reaction to AI\u2011led creative experiments has sometimes been immediate and caustic. A high\u2011profile example came when a major brand released an AI\u2011generated holiday advertisement and met widespread criticism for uncanny faces and an overall lack of warmth; journalists and critics framed the backlash as a symptom of rushed technical substitution for human craft. The brand defended the work as a collaboration between human storytellers and generative tools, but the episode underscored the reputational risk companies face when automated processes produce work that audiences perceive as inauthentic.<\/p>\n
Against that polarised backdrop, many industry voices and trade reports advise a cautious, case\u2011by\u2011case adoption strategy. Studios are experimenting where the economics and ethics feel manageable, while unions and regulators press for contractual safeguards, transparency about AI use and limits on how training datasets are created. Practitioners who have embraced hybrid workflows tend to stress three practical rules: use AI to augment repetitive or technically prohibitive tasks, keep humans in the loop for creative decisions and disclosure, and curate training data so models reflect the project\u2019s aesthetic rather than generic mass content.<\/p>\n
Looking forward, technical and commercial trends point to broader adoption of real\u2011time interpolation and other frame\u2011level AI tools in distribution and playback \u2014 a development that could make hybrid workflows more scalable and less threatening if paired with clear labour protections. The industry\u2019s immediate test will be whether those safeguards are written into contracts and studio practice, and whether practitioners can make \u201caugmentation\u201d mean genuine enhancement of craft, not its cheapening. As Bloomfield\u2019s example shows, the most defensible position for creatives is to treat AI as a specialised tool that can amplify hand work \u2014 not as a substitute for the labour and judgment that give animation its human resonance.<\/p>\n