全球最大独立 AI 原生影视公司,要打造「AI 版 Disney」World's largest independent AI-native film and TV company aims to build an 'AI version of Disney'


AI free token/AI免费积分

凭《怪物史莱克》取得奥斯卡最佳改编剧本提名的编剧 Joe Stillman、《冰雪奇缘》的选角导演 Scott Muller、《鬼妈妈》的艺术指导 John Lee,挑选在一部由全球最大的独立 AI 原生影视公司 Utopai Studios 主导的动画长片里相聚。

据《Variety》报道,这部预计于 2027 年暑期上映的家庭动画长片《The Most Serious Fart》,围绕一个格外严肃的屁 Siegfried 展开,他希望改变外界对自己的看法,并在旅途中学习接纳自己、找到归属感。

Image

过去三年,生成式 AI 讲给影视业听的故事大多关于替代,关键词是成本骤降、一人剧组、人人都是导演。

可下面要讲的是一个其作品即将登上院线的的 AI 影视公司——Utopai Studios 的故事。

它进了好莱坞工会,进了院线,巴西和德国的电视台想按普通剧集的售价买它的片子。

Utopai Studios 的首席运营官 Kevin Chong 坐在 TechRadar 记者对面时,遇到这样一个提问:「观众开始抵触 AI 电影了,购票之前会先查一部电影有没有用 AI,你们怎么办。」

他回答说:「Utopai 不把 AI 电影当作一个单独的片种,AI 只是一种工具,和视觉特效、虚拟制作没有本质区别。」

今年九月,Utopai 的定制视频模型 Utopai X,在 AA 的文生视频盲评榜上排到全球第二、全美第一,是榜单运行以来第一家凭自有模型跻身前列的影视公司。

Image

一镜一镜,AI 如何进入长片制作

AI 影视大致存在两种商业路径。

一种是工具生意,把模型做成订阅服务卖给创作者,按生成时长计费。

另一种是内容生意,公司自己承担开发和制作风险,把作品送进影视行业原有的交易平台,靠版权预售、电视播映权、流媒体、院线以及后续 IP 收入回收成本。

传统买家购买一部作品时,并不会因为工艺路线新鲜就降低风险标准。他们仍然要确认版权是否清晰,制作能不能按时完成,角色和画面能不能在几百个镜头中保持稳定,演员权益有没有后患,内容能不能符合当地播出规范,最后还要判断,观众究竟愿不愿意看。

所以,过去不少 AI 内容的难题都是走不到采购桌前。

Utopai 打破了这个僵局。

《Space Nation》是一部 8 集科幻剧集,每集约 45 至 60 分钟,由 Martin Weisz 执导全季。官方披露表明,项目已经进入制作,并计划按照好莱坞公会框架推进。

Image

《Space Nation》也是 Utopai 第一次在工会体系内完整走完一部剧集的生产流程。

作为首家加入好莱坞相关工会的 AI 影视公司,Utopai 在推进这部剧集制作的过程中,就 AI 内容的应用规则、真人演员与工作人员的参与比例、AI 辅助环节的边界等一系列难题,与工会维度展开了多轮协商。

Forbes 披露,Utopai 已经把《Space Nation》和历史史诗电影《Cortés》的部分国际发行权预售给涵盖巴西 Globo 体系、德国电视台在内的传统买家,而且使用的,是非 AI 内容已有的正常市面售价。

如果项目按计划交付,《Space Nation》和《Cortés》的预售收入合计最高可能达到约 1.1 亿美元。

Image

后者以 1519 年西班牙人进入阿兹特克帝国后的文明冲突为背景,剧本由 Nicholas Kazan 创作,预计在 2027 年全院线发行。Utopai 试图让真人表演与数字环境结合,用 AI 承担一部分过去需要大型置景、后期视效和现场执行完成的工作。

作为中国观众,惊喜的是,Utopai 决定与华策影视联合开发《西游记:消失的 500 年》,从孙悟空被压五行山的空白期重新寻找故事。

在 2027 年即将上映的作品中,还有由柏林银熊奖导演 Hyo-joo Yang 执导的亲情剧情片《Half Moon》。

Image

影片聚焦一个生活在德国的跨文化家庭,在家中被忽视的少女 Yeri,与自我放逐、以护理工作为生并背负着过往身心创伤的姨妈 Ah-Jin,在一个夏天里学会了理解与尊重彼此。

主体拍摄已于 2026 年 8 月在德国开机,影片计划于 2027 年上映。PAI 将为部分视觉元素给出兼容,例如一些游泳相关的实景段落,剧组先完成真人表演和实景拍摄,再用 AI 补全环境中那些依靠传统置景和特效也很难完成的部分。

PAI (Production Assistive Intelligence)是 Utopai 为了应对长片生产的连续性开发的制片智能平台。创作者可以先外观一个角色模型,再把这个角色放进多个情形,挑选机位,修改表演和环境,并在调整后重新生成画面。

Image

剧本会被拆成情形和镜头,人物、地点与道具参考被放进 Production Library;每个生成版本与原镜头关联,团队可以比较不同 take,保留修改历史,再把选中的素材送入 Adobe Premiere 或 DaVinci Resolve。

Image

给出视频生成能力的底层模型是 Utopai X,它在 MiniMax H3 基础上经过针对影视制作情形的专项训练,在光线反射、焦散成效、空间理解和跨镜头连贯性上有针对性的优化。

依据 Artificial Analysis,Utopai X 在音频同步与物理表现两个能力类别中排名第一,在光照与材质、摄影机控制两个类别中排名第二。

Image

当电影制作人挑选使用 Utopai X 时,PAI 会在该环境中生成视频内容。这两个角色是相互关联的:PAI 负责组织相关的工作和信息,而模型则负责生成视频素材。

这也解释了,为什么这些院线电影没有直接用很多耳熟能详的视频模型,反倒先用 Utopai X 融入自己的制作流程。

前段时间 Utopai 还宣布与 DeNA 合作,这是 PAI 在海外的商业落地项目之一。

Image

日本动画具备一套高度成熟且相对封闭的生产体系,许多工作室在工序管理上接近手工业式的精细操作,对外来工具的引入历来谨慎。

DeNA 挑选引入 PAI,很大程度上也是基于商业的考量:公司积累了大量内容资产,过去因制作成本高企长期处于冻结状态。

按照合作框架,Utopai 的 AI 制作专家团队配合 DeNA 的创意和制作人员,参与视觉概念开发、角色与环境外观、情形规划和迭代流程,但最终的故事、视觉风格、表演方式和总体基调,决定权留在 DeNA 的艺术家手里。

AI 原生影视,要把创作决定权交还给创作者

看完 Utopai 公布的片单,APPSO 发现它们恰恰对应了 AI 原生影视的几种形态。

其中一类是当下 AI 工艺最容易发挥、观众接受度最高的动画类内容。

Image

一类是像《Half Moon》这样的文艺片,只用 AI 辅助完成少量难以达成的视觉元素,核心仍然是真人表演和导演主导的叙事。

像《Cortés》《Space Nation》这类制作门槛极高、常规实拍方式很难完整展现的大制作故事内容,AI 在这里主要用来补全场面调度和视效上原本不可能完成的部分。

AI Native 已经成了各个行业拥抱 AI 的标签,在影视行业,谈到「AI 原生影视」,通常被认为演员由 AI 生成,情形由 AI 生成,剪辑、调色、合成也交给 AI,真人团队几乎不参与,剧组只是在给模型挑选素材和参数。

这种定义将 AI 和真人放到了对立面,也不利于让更好的故事走向观众。

在 APPSO 看来,AI 原生影视并非把人从制作流程里拿走,应该是用 AI 重写长片生产的基础设施,让原本拍不起、拍不快、拍不到的部分变得可拍,而叙事、表演、审美和最终责任仍然由懂故事的人承担。

笔者在 Utopai 身上看到了AI 原生影视的一个参考样本,这种模式给出了新的理解。

因而,AI 原生影视不等于「全 AI 电影」,它更像一套新的生产方式。

引入 AI 后,导演和编剧可以在前期更早地看见故事的空间关系、镜头节奏和视觉尺度,把原本停留在文字和口头描述里的想象,迅速转化为可讨论、可修改的预演版本。

在具体制作中,AI 更合适承担那些重复、庞杂、危险,或者传统方式很难扩展的工作。它可以协助搭建并不存在的城市,生成大规模群众场面,补齐复杂的环境变化,处理高风险动作的预演,也可以在后期依据导演已经确定的构图和光线要求,完成大量素材的迭代。

Image

这样一来,剧组的工作从「亲手完成每一个元素」转向「定义规则、监督过程、筛选结果」。

具体的决定贯穿从开发、选角、表演、镜头外观到剪辑定稿的每一个环节。

Image

随着 AI 深入影视制作,演员的表演权、创作者的署名权、训练素材的来源,以及生成内容的版权边界,也会成为制作流程的一部分。

将来的制作标准,既要判断画面是否逼真,也要追问它从何而来、由谁授权、经过谁的挑选,最终由谁承担责任。

从行业角度看,这种路径的意义也不只是降低成本。过去只有少数公司可以承担的宏大世界、复杂调度和跨地域拍摄,将来可能被更多具备好故事和明确创作意图的团队尝试。

制作门槛下降之后,竞争反而更集中在故事判断、导演能力、组织效率和作品的独特气质上。当「做出一个画面」不再稀缺,可以让观众相信这个画面为什么存在,便成为更稀缺的能力。

有生命力的 AI 原生影视,应该让导演更早看见自己的想法,让编剧具备更多验证故事的机会,让演员和美术、摄影、剪辑等创作者保留对作品的参与和表达,也让观众可以感受到每一个镜头背后的挑选。

如果一部电影最后只剩下模型的痕迹,那它只是一次工艺演示,只有让观众记住的是人物、情绪和那个世界,它才算电影。

所谓 AI 原生影视,终点从来不是让机器接管片场,从默片时代到 AI 时代,工艺的价值都应该是让更多好故事有机会走进电影院。

本文来源:https://www.ifanr.com/1683309?utm_source=rss&utm_medium=rss&utm_campaign=
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。

Screenwriter Joe Stillman, who earned an Oscar nomination for Best Adapted Screenplay for *Shrek*, *Frozen* casting director Scott Muller, and *Coraline* production designer John Lee have come together for an animated feature led by Utopai Studios, the world's largest independent AI-native film and television company.

According to *Variety*, the family animated feature *The Most Serious Fart*, expected to hit theaters in summer 2027, follows an exceptionally serious fart named Siegfried, who hopes to change how others see him and learns to accept himself and find a sense of belonging on his journey.

Image

Over the past three years, the stories generative AI has told the film and television industry have mostly been about replacement—the key themes being plummeting costs, one-person crews, and everyone becoming a director.

But the story ahead is about Utopai Studios, an AI film company whose work is about to hit movie theaters.

It has joined Hollywood unions, entered theatrical distribution, and TV networks in Brazil and Germany want to buy its content at the same rates as regular live-action productions.

When Utopai Studios COO Kevin Chong sat across from a TechRadar journalist and was asked, 'Audiences are starting to resist AI movies—they'll check before buying tickets whether a film used AI. What do you do?' he replied: 'Utopai does not treat AI films as a separate genre. AI is just a tool, no different in essence from visual effects or virtual production.'

This September, Utopai's custom video model Utopai X ranked second globally and first in the United States on AA's blind text-to-video evaluation leaderboard—the first film and television company to reach the top ranks relying solely on its own model since the leaderboard's inception.

Image

Frame by Frame: How AI Is Entering Feature-Length Production

There are roughly two commercial paths for AI-generated film and television.

One is a tools business: selling models as subscription services to creators, billed by generation time.

The other is a content business: the company itself assumes development and production risk, delivering works into the industry's existing distribution channels and recouping costs through pre-sales, television broadcasting rights, streaming, theatrical release, and subsequent IP revenue.

Traditional buyers do not lower their risk standards simply because a production technique is novel. They still need to confirm that copyrights are clear, that production can be completed on time, that characters and visuals remain consistent across hundreds of shots, that talent rights pose no future complications, that content complies with local broadcasting regulations, and—ultimately—whether audiences actually want to watch it.

So in the past, many AI content projects never made it to the procurement table.

Utopai has broken this deadlock.

*Space Nation* is an eight-episode sci-fi series, each episode approximately 45 to 60 minutes long, directed for the full season by Martin Weisz. Official disclosures confirm the project is in production and plans to proceed within the Hollywood guild framework.

Image

*Space Nation* is also Utopai's first complete run-through of a full series production process within the union system.

As the first AI film and television company to join relevant Hollywood unions, Utopai has engaged in multiple rounds of negotiations with the unions over a series of thorny issues—including rules for AI content application, the proportion of human actors and staff involvement, and the boundaries of AI-assisted workflows.

According to *Forbes*, Utopai has pre-sold partial international distribution rights for *Space Nation* and the historical epic *Cortés* to traditional buyers including Brazil's Globo network and German television stations, using the same standard market rates already established for non-AI content.

If the projects are delivered as planned, combined pre-sale revenue for *Space Nation* and *Cortés* could reach up to approximately $110 million.

Image

The latter is set against the backdrop of cultural clashes after the Spanish entered the Aztec Empire in 1519, with a screenplay by Nicholas Kazan, and is expected for a full theatrical release in 2027. Utopai is attempting to combine live-action performances with digital environments, using AI to take over some of the work that previously required large-scale sets, post-production VFX, and on-site execution.

Image

To Chinese audiences, the surprise comes from Utopai's decision to co-develop Journey to the West: The Lost 500 Years with Huace Film & TV, reviving a story from the blank period when Sun Wukong was trapped beneath the Five Elements Mountain.

Image

Among the works set to premiere in 2027 is the family drama Half Moon, directed by Berlin Silver Bear winner Hyo-joo Yang.

Image

The film focuses on a cross-cultural family living in Germany. Yeri, a teenage girl overlooked at home, and her aunt Ah-Jin, who lives in self-imposed exile, earns a living as a nurse, and carries physical and emotional scars from her past, learn to understand and respect one another over the course of a single summer.

Image

Principal photography commenced in Germany in August 2026, with the film slated for a 2027 release. PAI will provide compositing for select visual elements—for example, in some live-action swimming sequences, the crew first shoots the live-action performance and practical locations, then uses AI to fill in portions of the environment that would be extremely difficult to achieve with traditional set construction and visual effects.

Image

PAI (Production Assistive Intelligence) is Utopai's production intelligence platform developed to address the continuity challenges of feature-length filmmaking. Creators can first design a character model, place that character in various scenarios, select camera angles, adjust performances and environments, and then regenerate the imagery after modifications.

Image

In specific production work, AI is better suited to handle tasks that are repetitive, complex, dangerous, or difficult to scale with traditional methods. It can assist in building non-existent cities, generating large-scale crowd scenes, filling in complex environmental changes, handling previews of high-risk stunts, and completing iterations of vast amounts of footage in post-production based on the director's finalized composition and lighting requirements.

Image

Concrete decisions permeate every stage—from development and casting to performance, lens aesthetics, and final editing.

Image

Source: https://www.ifanr.com/1683309?utm_source=rss&utm_medium=rss&utm_campaign= If there are copyright issues, please notify us at 13951431913#139.com or submit a comment on the right.

暂无评论,快来发表第一条评论吧!No comments yet. Be the first!

📮 需求咨询📮 Inquiry