迪士尼想把你也拖进泥潭里。

内容来源:https://www.theverge.com/ai-artificial-intelligence/842992/disney-openai-sora-ai-slop-partnership
内容总结:
迪士尼与OpenAI近日宣布达成一项价值10亿美元的战略合作,引发业界广泛关注。根据协议,迪士尼将向OpenAI开放旗下包括皮克斯、漫威、星球大战等数百个经典角色的IP库,供其视频生成模型Sora的用户进行二次创作。部分生成内容将在迪士尼流媒体平台Disney+设立专区展示。
迪士尼首席执行官鲍勃·艾格将此次合作称为“行业重要里程碑”,强调其能为用户提供“更丰富、更个性化的角色互动体验”,并承诺在利用生成式AI拓展故事叙述边界的同时,“尊重并保护创作者及其作品”。然而,这项合作背后凸显出传统娱乐巨头与AI科技公司各取所需的现实考量:OpenAI借此获得重要资金注入,迪士尼则通过拥抱前沿技术展现其转型决心。
值得关注的是,迪士尼此前曾因AI公司未经授权使用其IP进行模型训练而发起诉讼,如今却主动付费开放核心IP库,这一转变颇具争议。尽管协议明确禁止AI模仿演员声音与肖像,但业界担忧这实则为迪士尼规避了向原版角色创作者支付报酬的义务。更现实的挑战在于,当前AI生成视频仍存在时长受限、逻辑混乱、质量不稳定等问题,难以支撑起专业影视内容需求。迪士尼计划初期聚焦用户生成的碎片化内容,但如何保障海量二创内容不偏离角色原初精神、避免出现不当演绎,仍存在巨大管理风险。
历史经验已敲响警钟:此前《堡垒之夜》游戏中由AI驱动的达斯·维德角色曾遭玩家诱导发表不当言论,而Sora社交平台上已出现知名人物被恶意篡改的视频案例。尽管OpenAI声称已建立内容审核机制,但在用户创作热情与监管成本的平衡中,迪士尼经典IP或将面临被海量低质、怪异二创内容稀释品牌价值的风险。
此次合作折射出娱乐产业在AI浪潮中的矛盾处境:一方面急切寻求技术赋能与流量新入口,另一方面尚未找到艺术创作与算法生成之间的有效平衡点。迪士尼能否真正实现“保护创作者”的承诺,抑或将沦为利用用户免费创作填充内容库的捷径,仍需时间检验。
中文翻译:
迪士尼与OpenAI新近达成的10亿美元合作,恰如一面镜子映照出我们所处的糟糕时代。作为获取这家生成式人工智能公司API及ChatGPT等工具的交换,迪士尼计划允许OpenAI的Sora AI视频生成器用户使用数百个迪士尼旗下角色创作短片。Sora AI用户将能随心所欲地生成大量诡异怪诞的皮克斯/漫威/星球大战风格内容,而迪士尼则会在Disney+平台设立专区展示部分作品。这场合作的走向可谓一目了然。
迪士尼正试图将你拖入内容泥潭
Disney+即将被毫无灵感的垃圾内容填满。
迪士尼首席执行官鲍勃·艾格将此次合作称为"行业重要时刻",声称这将为迪士尼客户提供"更丰富、更个性化的方式与他们热爱的迪士尼角色和故事建立连接"。他还表示,这笔交易旨在"通过生成式AI拓展我们叙事的影响力,同时尊重并保护创作者及其作品"。
这或许确实是艾格对与OpenAI联姻前景的真诚构想,但只要客观审视Sora AI的本质,以及迪士尼过往在生成式AI领域的实验轨迹,便能清晰预见这场合作将以极其荒唐的方式收场。
究其本质,此次合作是OpenAI获取新一轮资本注入的手段,也是迪士尼对争夺科技领域主导权的AI巨头做出的强势站队。迪士尼承诺的10亿美元虽不能彻底解决OpenAI的现金流困境,却使其成为首家与老牌娱乐工作室达成如此规模合作的AI企业。尽管Sora不会立即整合进迪士尼项目制作流程,却为工作室向公众展示其拥抱生成式AI的意向提供了具体途径。而从另一角度看,迪士尼将获得海量可即时投放的劣质内容,且无需支付任何创作报酬。
在迪士尼等多家工作室因AI公司使用未经授权的IP训练模型而提起诉讼数月后,迪士尼竟转而向OpenAI付费使用自家旗下复仇者联盟、曼达洛人等角色,此举着实令人费解。表面看来这笔交易明显失衡,而当我们审视Sora AI的实际能力时,整件事更显荒诞。
尽管Sora能将几句文字转化为光鲜的视频片段,但AI生成的短片始终存在时长受限、画面诡异、逻辑矛盾等缺陷,难以匹配观众愿意付费购买的主流娱乐内容标准。这正是多数生成式AI制作的"电影"和广告沦为视觉垃圾的原因。但或许正是这些局限性,促使迪士尼决定从碎片化的用户生成内容领域试水。
粉丝通过二次创作解构企业IP的行为本非新鲜事,但迪士尼此次的特别之处在于其商业化意图。Sora AI的整合将使迪士尼直接获取内容输送渠道,而订阅用户反而需向工作室付费使用该工具。这套体系实质上将Disney+用户转化为既是消费者又是无偿劳工的复合体——这种场景想必深得迪士尼管理层青睐。然而回顾迪士尼上次开放IP改编权限的经历,事态正是以最可预料的方式迅速失控。
尽管玩家无法让《堡垒之夜》中AI驱动的达斯·维达NPC为所欲为,但通过反复诱导仍能迫使该角色用詹姆斯·厄尔·琼斯的嗓音说出充满仇恨的偏执言论。Epic Games虽采取措施限制维达的异常行为,但玩家总能找到漏洞——这恰恰是人性使然。同样"因为能做所以去做"的冲动,正是Sora AI专属社交视频平台上诸多短片的创作动力,在那里不难找到罗杰斯先生、鲍勃·罗斯等形象说着冒犯性台词做着出格行为的视频。
OpenAI或许向迪士尼管理层承诺已建立防护机制,防止《超人总动员》角色高喊种族歧视言论等情况。但即便工作室审核每段涉及迪士尼角色的Sora AI产出,仍将面临源源不断的劣质内容——这些作品与IP原初创造者(技艺精湛的人类艺术家)期望的品质相去甚远。
艾格所谓"合作旨在尊重保护成就迪士尼今日地位的艺术家与艺术作品"的保证实在难以取信于人。迪士尼声称Sora AI不得复制任何角色相关演员的声纹或肖像,但这恰恰暴露了整个项目的空洞本质。此举既确保迪士尼无需重新酬谢那些赋予角色灵魂的表演者,也暴露出迪士尼认为观众仅凭熟悉却脱离语境的面孔就能获得娱乐的傲慢认知。
热门聚焦
- AI企业渴求新型互联网生态,宣称已掌握关键密钥
- 谷歌正研发实验性浏览器与革命性网页应用架构
- GPT-5.2成为OpenAI智能体竞赛的最新战略部署
- ChatGPT"成人模式"预计2026年首季亮相
- Ayaneo首款安卓手机重塑Xperia Play经典形态
英文来源:
Disney and OpenAI’s new $1 billion partnership feels emblematic of the terrible times we’re living through. In exchange for access to the generative AI firm’s APIs and tools like ChatGPT, the studio plans to let users of OpenAI’s Sora AI video generator create clips featuring hundreds of Disney-owned characters. Sora AI users will be able to generate as much uncanny Pixar / Marvel / Star Wars slop for themselves as they want, and Disney will share some of it in a special section on the Disney Plus platform. It’s not hard to see where this is going to go.
Disney wants to drag you into the slop
Disney Plus is about to become filled with uninspired garbage.
Disney wants to drag you into the slop
Disney Plus is about to become filled with uninspired garbage.
Disney CEO Bob Iger described the partnership as marking “an important moment for our industry” and said that it will give the studio’s customers “richer and more personal ways to connect with the Disney characters and stories they love.” Iger also said that the deal is meant to “extend the reach of our storytelling through generative AI, while respecting and protecting creators and their works.”
That might genuinely be how Iger thinks getting into bed with OpenAI will play out, but when you take an honest look at what Sora AI is and how Disney’s past experiments with generative AI have gone, it’s clear that this is all going to end very, very stupidly.
At its core, this partnership is about OpenAI securing a new influx of capital and Disney coming out in strong support of one of the big gen AI companies fighting for tech sector dominance. The $1 billion Disney has pledged won’t completely solve OpenAI’s cash flow problems, but it will make it the first company of its kind to do business with a legacy entertainment studio on this scale. While Sora isn’t immediately being integrated into the production workflows for Disney’s projects, it will give the studio a concrete way of signaling to the public its interest in using generative AI. And on the flip side of things, Disney is getting access to a boatload of slop content it can stream without having to pay anyone to make it.
After months of various studios including Disney filing lawsuits against gen AI companies for building models trained on unlicensed IP, it seems beyond odd that Disney would pay OpenAI for its use of Disney’s characters like the Avengers and the Mandalorian. On its face, the transaction reads as being pointedly imbalanced, but the situation is even stranger when you look at what Sora AI is capable of.
Though Sora can turn a couple of text sentences into glossy video footage, AI generated clips are always short, strange, and filled with inconsistencies that make them ill-suited to be used in the kinds of entertainment people are typically willing to pay for. That’s why so many “movies” and ads created with generative AI look like absolute garbage. But those limitations are also probably the reason that Disney plans to begin with a focus on bite-sized, user-generated content.
Fans entertaining themselves by riffing on corporate-owned IP with unique creations of their own is nothing new, but what’s different here is the way that Disney wants to capitalize on that behavior. Sora AI integration would give Disney direct access to a content pipeline that subscribers would be paying the studio to use. This system would effectively invite Disney Plus users to become customers and unpaid employees, which is a scenario that probably sounds ideal to Disney’s executive leadership. But the last time Disney gave people the ability to mess around with its IP this way, things quickly devolved in the most expected way possible.
While players couldn’t make Fortnite’s gen AI-powered Darth Vader NPC do whatever they wanted, it was possible to coax the character into saying hateful, bigoted things in James Earl Jones’ voice with enough prodding. Epic Games took some measures to try to keep Fortnite’s Vader from behaving in inappropriate ways, but — shocker — players still found workarounds because that’s the way people are. That same “let’s do it because we can” energy is what animates a lot of the clips you see on the dedicated Sora AI social video platform, where it’s easy to find footage of figures like Mister Rogers and Bob Ross saying and doing offensive things.
OpenAI has likely told Disney leadership that it has put measures in place to keep people from, say, making characters from The Incredibles shout racist invective. But even if the studio reviews every single Sora AI output involving Disney characters, there is still going to be an endless stream of content that’s far from the quality the IP’s original creators — skilled human artists — intended.
It’s hard to believe Iger’s assurances that this partnership was made with the intention of respecting and protecting the art and artists that have made Disney such a powerful company. The company says that Sora AI will not be allowed to reproduce the voices or likenesses of any of the actors associated with its characters, but that, too, feels like a sign of how soulless this entire endeavor is. It gives Disney a solid way of making sure that it does not have to re-pay the performers whose work originally brought these characters to life. It also makes it seem like Disney thinks that people can be entertained simply by being shown familiar, decontextualized faces.
Most Popular
- AI companies want a new internet — and they think they’ve found the key
- Google is building an experimental new browser and a new kind of web app
- GPT-5.2 is OpenAI’s latest move in the agentic AI battle
- ChatGPT’s ‘adult mode’ is expected to debut in Q1 2026
- Ayaneo’s first Android phone looks like a return of the Xperia Play