好莱坞对人工智能束手无策。
内容总结：
【硅谷与好莱坞在AI浪潮下的激烈碰撞】当OpenAI在开发者大会上将视频生成模型Sora誉为“献给创作者的礼物”时，好莱坞制片厂的高管们正在洛杉矶的行业论坛上为AI的冲击焦灼不安。这场科技与娱乐的对话，暴露出两个世界难以弥合的分歧。
在旧金山的媒体问答环节，OpenAI首席执行官萨姆·奥尔特曼宣称创作者对Sora的潜力“充满热情”，并将其类比为“新一代同人创作”。然而仅隔一日，洛杉矶 Bloomberg 论坛上的影视巨头们却展现出截然不同的态度。尽管与会者反复强调“重视版权”，却无人直接回应OpenAI未经许可使用受版权保护内容训练模型的核心争议。
值得注意的是行业应对策略的分化：网飞联合CEO格雷格·彼得斯在谈及Sora时避重就轻，派拉蒙CEO大卫·艾里森则将AI温和地称为“新型创作铅笔”。唯有华纳音乐CEO罗伯特·肯塞尔明确表态，要求AI公司必须获得授权才能使用其内容训练模型，并警告违规者将承担后果。
这种分野背后是行业特性的深刻差异。音乐行业因曾应对流媒体冲击，已形成更成熟的集体维权机制。而好莱坞至今未能形成统一应对策略，这种迟疑可能使科技公司继续沿用“先侵权后道歉”的既定套路。正如业内人士所言，OpenAI的训练方式绝非偶然，而是科技行业实现市场主导的经典策略重演。
（根据科技记者亚历克斯·希斯专栏内容整理）
中文翻译：
本文节选自《科技行业与人工智能通讯》作者亚历克斯·希斯的专栏，该内容每周专为The Verge订户同步刊发。
好莱坞对人工智能束手无策
OpenAI正在以游戏速通模式席卷娱乐产业。
OpenAI正在以游戏速通模式席卷娱乐产业。
本周我亲眼见证了硅谷与好莱坞在应对AI问题上存在的巨大分歧。
在OpenAI开发者大会上，萨姆·奥尔特曼将新款Sora应用作为献给内容创作者的礼物。他反而暗示，OpenAI目前限制用户制作更多类型的AI视频，审查标准可能过于严苛。
"总体而言，创作者、版权持有者都对这项技术的潜力感到振奋。"奥尔特曼在周一旧金山媒体问答会上表示，"人们相信这将深化情感联结，堪称新一代的同人创作。"
次日我出席洛杉矶举办的彭博社Screentime论坛时，听到媒体高管、经纪公司和影业负责人将AI比作迎面袭来的陨石。此时Sora在应用商店的下载量刚突破百万，成为全场焦点。我离场时的深刻感受是：好莱坞领袖们对AI带来的威胁毫无应对之策，这项技术的演进速度已超出他们的理解范围，终将被时代洪流碾过。
派拉蒙天舞影业CEO大卫·埃里森将AI称为创造力的"新型画笔"
论坛上"我们重视版权"这类表态如同祷告词般被反复诵念，次数多到令人麻木。但与此同时，所有与会者都刻意回避一个事实：OpenAI显然在未经授权的情况下使用他们的知识产权进行训练，并且推出的产品至少在初期毫不掩饰这一点。好莱坞领军者既未能就此问题形成统一立场，更未提出具体应对方案，这值得所有从业者警惕。
网飞联合CEO格雷格·彼得斯在台上完全回避了关于Sora的提问，转而大谈AI在制片各环节中那些乏善可陈的应用。派拉蒙天舞影业CEO大卫·埃里森也选择强调AI作为工具的无争议属性，将其喻为创造力的"新型画笔"。唯一接近触及问题核心的是华纳音乐CEO罗伯特·肯塞尔，他明确表示训练AI必须获得华纳内容授权，违规者将承担后果。
音乐行业能比那些对AI瞻前顾后的经纪公司形成更清晰的立场并不意外。这些唱片公司曾成功应对音乐流媒体变革，如今作为 consolidated 利益共同体，更有底气与AI企业博弈。肯塞尔甚至预测AI将长期利好音乐产业，就像YouTube最终解决版权争议后，成长为娱乐业重要分发平台那样。
这个判断或许适用于音乐领域，但好莱坞其他环节的集体失语，意味着AI公司将继续奉行"先斩后奏"的策略。OpenAI选择以此方式训练Sora绝非偶然，而是彰显其吞噬所有内容喂养AI的肆无忌惮。奥尔特曼不过是在沿用科技界惯用的垄断套路，这次又有谁能指责他呢？
英文来源：
This is an excerpt of Sources by Alex Heath, a newsletter about AI and the tech industry, syndicated just for The Verge subscribers once a week.
Hollywood has no idea what to do about AI
OpenAI is speedrunning the entertainment industry.
OpenAI is speedrunning the entertainment industry.
This week, I got an up-close look at how far apart Silicon Valley and Hollywood are on what to do about AI.
First, at OpenAI DevDay, Sam Altman presented the new Sora app as a gift to content creators. If anything, he suggested, OpenAI was being too censorious by not allowing people to make even more kinds of AI videos.
“On the whole, creators, rights holders, people are very excited about the potential of this,” Altman said during a media Q&A in San Francisco on Monday that I attended. “They believe it will deepen connection. It’s kind of like a new generation of fanfiction.”
The next day, I arrived at Bloomberg’s Screentime event in Los Angeles to hear how media executives, agents, and studio heads felt about the AI meteor that’s heading their way. Sora had just hit 1 million downloads in the App Store and was top of mind for everyone. Altogether, I came away with the impression that Hollywood’s leaders still have no idea what to do about the risk AI poses, and they’re going to be steamrolled by technology that’s moving faster than they can comprehend.
Paramount Skydance CEO David Ellison called AI a “new pencil” to create with
I lost count of how many times a version of the phrase “we care about copyright” was invoked at Screentime like a prayer. At the same time, no one at the event wanted to specifically address the fact that OpenAI clearly trained on their IP without permission and unleashed a product that, at least initially, had no shame in making that clear. The fact that Hollywood’s leaders are unable to share a public perspective on this issue, or more importantly, what they are going to do about it, should be alarming to everyone working in the business.
While onstage, Netflix co-CEO Greg Peters completely dodged a question from Bloomberg’s Lucas Shaw about Sora specifically, and instead waxed about the more boring ways that AI is being used throughout nearly every part of the production process. Paramount Skydance CEO David Ellison also chose to tout the less controversial, tool-like aspect of AI, calling it a “new pencil” to create with. The only executive I heard come close to addressing the real issue on everyone’s minds was Warner Music CEO (and former YouTube exec) Robert Kyncl, who made clear that Warner’s content must be licensed to train on, and that there will be repercussions for those who don’t follow the rules.
It’s not surprising that the music industry has a stronger perspective than, say, the hemming and hawing about AI currently being done by the big talent agencies. The labels are better positioned to take on AI companies as a consolidated group of players who have navigated a version of this problem before with the rise of music streaming. Kyncl went so far as to predict that AI will benefit the music industry long-term, similar to how YouTube eventually solved its copyright problem and evolved into a major distribution platform for the entertainment industry.
He may be right about music specifically, but the lack of collective action from the rest of Hollywood means that the AI companies are poised to keep getting away with asking for forgiveness instead of permission. OpenAI’s decision to train Sora in this manner was a deliberate choice, not an accident, and it showed a complete lack of regard for the implications of sucking up everyone’s content to feed its AI. Altman is simply following the same playbook the tech industry has used in the past to achieve dominance, so who can blame him this time?