内容来源：https://www.theverge.com/ai-artificial-intelligence/932464/musk-v-altman-proved-that-ai-is-led-by-the-wrong-people

内容总结：

马斯克诉奥特曼案：一场信任崩塌的AI控制权之争

本周一，备受瞩目的“科技界年度审判”——马斯克诉奥特曼案在陪审团仅两小时审议后作出裁决，因诉讼时效问题驳回马斯克全部诉求。这场持续三周的法律较量表面无果而终，却暴露出人工智能行业更深的信任危机：掌控万亿美元产业的顶尖人物，似乎都难以让人信任。

案件核心围绕OpenAI控制权展开。马斯克指控其联合创始人奥特曼不应主导AI未来方向，奥特曼律师则质疑马斯克本人诚信。法庭证据显示，OpenAI创立初衷正是为了防止强人工智能落入“错误之人”手中。联合创始人格雷格·布罗克曼和伊利亚·苏茨克维曾强烈反对个人独揽大权，甚至不惜破坏一项可能让马斯克成为“AI独裁者”的巨额交易。

2023年11月，OpenAI董事会突然罢免奥特曼CEO职务（即“闪变事件”），成为庭审焦点。苏茨克维耗时一年起草52页备忘录，指控奥特曼“持续说谎、破坏高管关系”。时任CTO米拉·穆拉蒂更作证称，奥特曼曾对其宣称法律团队已批准跳过某模型的安全审查，而该说法实为虚假。

马斯克律师在结案陈词中强调，多名与奥特曼共事多年的证人当庭指证其说谎，“如果你们无法信任他，被告就无法获胜”。但马斯克自身形象同样受损。现任OpenAI首席未来学家约书亚·阿奇姆作证称，马斯克为与谷歌竞争采取“明显危险且鲁莽”的AGI研发路径。其旗下xAI同为营利机构，安全措施却相当松懈。OpenAI律师犀利指出，马斯克要求OpenAI保持“开放”，实则痴迷于“对AGI的统治权”。

庭上证词还揭示更多细节：穆拉蒂曾参与罢免奥特曼，后又转向支持其复职，却对自身角色“毫无兴趣”透露；马斯克密友希冯·齐里斯在担任OpenAI董事时，曾询问“是否应保持亲近以获取信息”，却隐瞒已为其生育两名子女；布罗克曼承认，若OpenAI在马斯克未参与情况下转为营利，马斯克“完全可以说我们对他不诚实”。

皮尤研究中心调查显示，美国民众对AI担忧已达历史峰值：半数成年人表示“AI日常使用增加令人担忧多于兴奋”，近60%认为对AI应用“几乎没有控制权”。多地爆发反对数据中心建设抗议活动，甚至出现针对奥特曼住宅的袭击企图。多位科技CEO私下承认已准备末日地堡等应急计划。

庭审中一份2015年邮件引发深思：奥特曼和马斯克曾联合起草信件，提议成立“AI安全监管机构”，致信微软CEO萨提亚·纳德拉请求支持。纳德拉回复拒绝，认为应由企业推动“联邦资金和研究”，而非监管。奥特曼随即同意修改信件，将监管选项推迟至“假如和当”需要时。这一细节被观察者视为行业“口惠而实不至”的缩影。

社交媒体调侃称：“如果不可信度有质量，把马斯克和奥特曼放在一起，法庭和整个地球都会坍缩成黑洞。”马斯克在X平台表示将提起上诉，OpenAI未作回应。这场审判最终指向一个令人不安的结论：那些被寄予厚望守护AI未来的领导者，或许正是最需要被监督的对象。

中文翻译：

今年的科技行业诉讼大案——马斯克诉奥特曼案，归根结底是一场控制权之争。埃隆·马斯克主张，他与萨姆·奥特曼共同创立的如今规模庞大的OpenAI公司，不应由后者来主导人工智能的未来方向。而奥特曼的律师团队则反过来质疑马斯克本人的可信度。陪审团在经过仅两小时的商议后，于周一作出裁决，以诉讼时效已过为由驳回了马斯克的诉求。

马斯克诉奥特曼案证明：人工智能正被错误的人掌控。

公众对人工智能行业的看法本已每况愈下。一连串不可信的高管更让情况雪上加霜。

从严格的法律意义上讲，三周的证词陈述最终毫无结果。但这场审判却揭示了一个更令人咂舌的深层结论：这场大戏中几乎没有人值得信任。科技界一些最有影响力的人物在性情上似乎根本无法诚实相待。倘若果真如此，便引出一个更大的问题：为何是他们掌控着一个即将颠覆人们生活的万亿美元产业？

在马斯克和奥特曼双方的证词中，OpenAI成立之初的目标都是为了防止强大的人工智能被错误的人所拥有和推动。证词和证据显示，其创始团队曾对谁将掌控通用人工智能（AGI）忧心忡忡——AGI是一个流行词，指的是在广义上等同或超越人类知识和能力的人工智能。他们极度忌惮谷歌DeepMind及其领导者德米斯·哈萨比斯。2015年，奥特曼曾表示，他一直在思考是否有什么能“阻止人类发展人工智能”——在得出“这不可能”的结论后，他希望“除了谷歌以外的其他方先做到这一点”。

联合创始人格雷格·布罗克曼和伊利亚·苏茨克维强烈反对由一人独掌控制权，他们甚至不惜破坏一笔利润丰厚的交易——用他们的话说，这笔交易可能会让马斯克获得“人工智能独裁”。在写给奥特曼的同一封邮件中，布罗克曼和苏茨克维质疑了他的动机，写道：“在此过程中我们始终无法完全信任你的判断……AGI真的是你的首要动机吗？它与你个人的政治目标有何关联？”

这些担忧很快便成了现实。马斯克诉奥特曼案的一个核心焦点是“那次事件”——2023年11月OpenAI董事会罢免奥特曼CEO职务的那五天。苏茨克维花了一年多时间策划他的下台，整理了一份长达52页的备忘录，指控奥特曼“一贯撒谎、削弱其他高管的威信，并挑拨他们相互对立”。此事的影响远不止高管内斗，还可能波及人工智能系统的公开部署。例如，时任首席技术官米拉·穆拉蒂在法庭上作证称，奥特曼曾告诉她，OpenAI的法律团队已同意跳过对某一款模型的安全审查——她说，这一说法后来被证实是虚假的。

在结案陈词中，马斯克的律师史蒂文·莫洛反复强调，有长长一串证人在宣誓后作证，称奥特曼在某种程度上是个撒谎者——而所有这些人都与奥特曼共事多年。“被告绝对需要你们相信萨姆·奥特曼，”莫洛对陪审团说，“如果你们不能信任他，如果你们不相信他，他们就赢不了。就这么简单。”

但在庭审过程中，如今领导其航天公司SpaceX旗下竞争实验室xAI的马斯克，表现也不见得好到哪里去。现任OpenAI首席未来学家约书亚·阿希姆作证称，马斯克与谷歌的竞赛导致他采取了“明显不安全且鲁莽”的方式来追求AGI。当他和其他人提出担忧时，马斯克争辩说OpenAI的营利化转型创造了忽视安全的激励机制——但他自己的xAI也是一家营利性公司，其对安全的态度充其量只是随意应付。而且，为了确保OpenAI保持“开放”，马斯克本人对控制它有着近乎偏执的渴望。在结案陈词中，OpenAI的一位律师萨拉·埃迪告诉陪审团，马斯克“想要的是对AGI的支配权”。

正如一位X用户所言：“如果‘不可信’有质量，把马斯克和奥特曼靠得太近，会让整间法庭乃至整个地球都坍缩成一个黑洞。”

OpenAI未立即回应置评请求。马斯克在X平台上发布声明称他将提起上诉。

问题并不只局限于马斯克和奥特曼。庭审证据显示，穆拉蒂曾帮助把奥特曼赶下台，随后又倒戈支持他复职，同时对于自己扮演的角色表现得“完全没兴趣”去披露。与马斯克关系密切、曾任职于OpenAI董事会的希冯·齐里斯在马斯克离开期间曾问他是否“更希望我与OpenAI保持亲近友好关系以维持信息流通”——却未透露当时她已与马斯克育有两个孩子。布罗克曼的日记在马斯克的案件中发挥了关键作用；他曾一度承认，如果OpenAI在马斯克未参与的情况下转向营利模式，马斯克可以“正确”地声称“我们对他不诚实”。

马斯克诉奥特曼案给了两人一个互相泼脏水的机会，理论上也让他们各自有机会将自己塑造成更严谨可靠的人工智能守护者。但一个更明显的结论是，人工智能行业里这些家喻户晓的名字，往好里说是天真幼稚，往坏里说则是虚伪至极，几乎不考虑自己行为的后果。

公众对人工智能的情绪正处于历史最低点。在去年夏天皮尤研究中心的一项调查中，半数美国成年人表示“日常生活中人工智能使用增加让他们感到担忧多于兴奋”——仅有10%的人表示兴奋多于担忧。这些担忧多与失业相关，但反对在全国范围内大规模建设数据中心的抗议活动也在激增。一些抵制行动甚至可能演变为暴力，据称有人曾两次试图袭击奥特曼的住所。许多科技公司CEO自己也承认，他们为情况万一糟糕至极准备了地下掩体或其他末日应对计划。

这些公司向公众传递的信息是：人工智能赋予用户力量。但皮尤研究中心2025年的一项研究发现，近六成美国成年人认为自己对人工智能如何在生活中被使用几乎没有或完全没有控制权。在美国，有意义的政府监管——至少能提供一定外部监督——的前景仍摇摆不定。而如今，人工智能领域最大玩家们为了维持控制会走多远，比以往任何时候都更加清晰。

在审判过程中涌现的大量证据中，有一份文件展示了奥特曼和马斯克愿意放弃部分权力的罕见例子。2015年3月，奥特曼给微软CEO萨提亚·纳德拉发了一封邮件，提出了一个简单的请求：签署他和马斯克正在起草的一封信，要求美国政府建立“一个人工智能安全新监管机构”，并应对“大多数人忽视的、人类存续面临的最大风险”。几周后，纳德拉回复了，否定了这个想法。“人类安全问题与控制难题终将成为现实议题，”他说。但他坚称，高管们应该呼吁的是“联邦资金投入和对研究的鼓励”，而非监管。奥特曼随即表示同意。他承诺，那封信将会修改——把监管人工智能行业的选项留到“如果且当”之时。

英文来源：

The tech trial of the year, Musk v. Altman, was ultimately a fight for control. Elon Musk argued that Sam Altman, with whom he helped found the now-massive company OpenAI, shouldn’t direct the future of AI. Altman’s lawyers, in turn, poked at Musk’s own credibility. A jury came to a verdict on Monday after just two hours of deliberation, dismissing Musk’s claims due to the statute of limitations.
Musk v. Altman proved that AI is led by the wrong people
Public opinion of the AI industry is already sinking. A parade of untrustworthy executives makes it look worse.
Musk v. Altman proved that AI is led by the wrong people
Public opinion of the AI industry is already sinking. A parade of untrustworthy executives makes it look worse.
In a strictly legal sense, three weeks of testimony added up to nothing. But the trial offered a more damning broader takeaway: Almost nobody in this saga seems worth trusting. Some of the most powerful people in tech seem temperamentally incapable of dealing with each other honestly. And if that’s true, it raises a bigger question: Why are they in control of a trillion-dollar industry that’s set to upend people’s lives?
OpenAI was, in the testimony of both Musk and Altman, founded to stop powerful AI from being owned and advanced by the wrong people. Testimony and evidence showed its founding team fretting about who would control artificial general intelligence (AGI), a buzzword for AI that broadly equals or surpasses human knowledge and ability. They deeply feared Google DeepMind and its leader, Demis Hassabis. In 2015, Altman said he’d been mulling over whether anything could “stop humanity from developing AI” — and after concluding it was impossible, that he wanted “someone other than google to do it first.”
Fellow cofounders Greg Brockman and Ilya Sutskever so strongly opposed one-person control that they seemed willing to torpedo a lucrative deal that could — in their words — give Musk an “AI dictatorship.” In a part of the same email addressed to Altman, Brockman and Sutskever questioned his motivations, writing, “We haven’t been able to fully trust your judgements throughout this process … Is AGI truly your primary motivation? How does it connect to your political goals?”
These concerns would be quickly borne out. A central focus of Musk v. Altman was “the blip,” a five-day period in November 2023 when OpenAI’s board removed Altman as CEO. Sutskever had spent more than a year architecting his ouster, assembling a 52-page memo alleging “a consistent pattern of lying, undermining his execs, and pitting his execs against one another.” The implications were broader than executive infighting, potentially impacting the public rollout of AI systems. Then-CTO Mira Murati, for instance, testified in court that Altman told her OpenAI’s legal team had okayed skipping a safety review for one of its models — a statement, she said, that turned out to be false.
In closing arguments, Musk attorney Steven Molo hammered home the long list of people who had testified under oath that Altman was, in one way or another, a liar — all of whom Altman had worked with for years. “The defendants absolutely need you to believe Sam Altman,” Molo told the jury. “If you cannot trust him, if you don’t believe him, they cannot win. It’s that simple.”
But during court proceedings, Musk — who now leads competing lab xAI, under his space company SpaceX — didn’t come off any better. Joshua Achiam, now OpenAI’s chief futurist, testified that Musk’s race against Google led him to take an “obviously unsafe and reckless” approach to achieving AGI. When he and others raised concerns, he says, Musk argued that OpenAI’s for-profit makeover created incentives to disregard safety, but his own xAI is for-profit and has, at best, a haphazard approach to safety. And in the name of making sure OpenAI remained open, Musk was obsessive in his need for control over it. In closing arguments, Sarah Eddy, one of OpenAI’s attorneys, told the jury that Musk “wanted dominion over AGI.”
As one X user put it, “if untrustworthyness had mass, putting Musk and Altman too close to one another would collapse the courtroom and all of earth into a black hole.”
OpenAI did not immediately respond to a request for comment. On X, Musk posted a statement saying he’d be filing an appeal.
It’s not just Musk and Altman, either. Trial evidence suggested Murati helped get Altman removed, then switched sides to support his reinstatement while appearing “totally uninterested” in disclosing the role she’d played. Shivon Zilis, a close Musk associate who served on OpenAI’s board, asked Musk if he’d “prefer I stay close and friendly to OpenAI to keep info flowing” during his departure — avoiding revealing that she had two children with him at the time. Brockman’s diary entries played a key role in Musk’s case; at one point, he admitted Musk could “correctly” claim “we weren’t honest with him” if OpenAI made a for-profit shift without his involvement.
Musk v. Altman gave each man an opportunity to sling dirt at the other and, in theory, establish himself as the more scrupulous guardian of AI. But a more obvious takeaway is that several of the AI industry’s household names are at best naive — and, at worst, hypocrites with little regard for the consequences of their actions.
Public sentiment about AI is at an all-time low. In a Pew Research survey from last summer, half of US adults said the “increased use of AI in daily life makes them feel more concerned than excited” — and only 10 percent said they felt more excited than concerned. Many of these concerns are related to job loss, but protests are also surging against mass data center construction across the country. Some resistance has turned potentially violent, with individuals allegedly attempting to attack Altman’s home on two occasions. And many tech CEOs themselves maintain that they have bunkers or other doomsday-prepping plans for if things go horribly wrong.
These companies push public messaging that AI empowers its users. But a 2025 Pew Research study found that nearly 60 percent of US adults feel they have little to no control over how AI is used in their lives. In the US, the prospect of meaningful government regulation — which could at least offer some level of external oversight — remains shaky. And now, it’s clearer than ever how far the AI world’s biggest players will go to maintain control.
Amid the trial’s reams of evidence, one document offers a rare example of Altman and Musk offering to cede some power. In March 2015, Altman emailed Microsoft CEO Satya Nadella with a simple request: Sign a letter that he and Musk were drafting, asking the US government to establish “a new regulatory agency for AI safety” and address “the biggest risk to the continued existence of humanity that most people are ignoring.” Weeks later, Nadella responded to shut down the idea. The “issue of human safety and the control problem will become real issues,” he said. But executives, he insisted, should be calling for “federal funding and encouragement of research,” not oversight. Altman promptly agreed. The letter, he promised, would be changed — leaving the option of regulating the AI industry “if and when.”