内容来源：https://www.wired.com/story/artificial-roasts-ais-creators-and-sends-a-dark-warning-about-its-dangers/

内容总结：

人工智能题材新片《Artificial》近日在纽约电影节首映，这部由卢卡·瓜达尼诺执导的黑喜剧聚焦OpenAI与全球人工智能军备竞赛，以讽刺笔触描绘了萨姆·奥尔特曼、伊利亚·苏茨克维尔、埃隆·马斯克等硅谷核心人物。影片首映前便风波不断——原定发行方亚马逊米高梅在亚马逊与OpenAI达成500亿美元合作后悄然弃片，独立片商Neon接盘。纽约电影节艺术总监丹尼斯·林在台上致辞时也隐晦提及了影片周边的“种种状况”。

影片对OpenAI及其利益相关者的刻画几乎毫无褒意，将他们塑造成虚荣、鲁莽且令人难以忍受地乏味的科技巨头。影评人不可避免地将本片与2010年大卫·芬奇的《社交网络》相比较——同样涉及硅谷题材，同样由安德鲁·加菲尔德主演——但两部影片的取向截然相反。《社交网络》冷峻精致，将程序员塑造成行业巨擘与未来缔造者；《Artificial》则近乎超现实地呈现为一出黑色喜剧，把科技大佬们拍成了自命不凡的书呆子。

加菲尔德饰演奥尔特曼，蹦蹦跳跳、少年气十足，奥尔特曼本人近日还称赞这一走路姿态“相当准确”。俄罗斯演员尤拉·鲍里索夫饰演苏茨克维尔，笨重地踱步，像极了一个怪异的弗兰肯斯坦式怪物。艾克·巴林霍尔茨饰演马斯克，大摇大摆、趾高气扬，如同阿尔法大猩猩般统领着一屋子电脑极客。

影片的核心张力在于：鲍里索夫饰演的这位心地善良、天性不会说谎的研究者恳求为AI设置安全护栏、放慢步伐，使其与人类更广泛的利益“对齐”；而主张技术与资本指数级增长的一方，几乎囊括了其余所有角色。影片甚至将“安全”本身也呈现为一门生意——对OpenAI潜在危害的担忧被包装成有利于公司底线的卖点：敲响警钟意味着透明与尽责，而危害能力本身则被当作技术进步的证明。赢了也是输，输了也是输。

《Artificial》对当前AI竞赛提出了一个令人信服却又令人沮丧的论点：技术已然存在，正如上世纪核武器一样，问题不在于力量本身，而在于谁被允许驾驭它、如何驾驭。当所有参与者遵循的规则各不相同时，伦理问题便沦为学术空谈。如果不是OpenAI掌控计算的未来，那就会是Anthropic、谷歌或马斯克——或者，最糟糕的情况，是中国。

对于普通观众而言——他们接触AI可能仅限于邮件润色或日程管理——很难判断影片的警示该被多当真，尤其片中几乎每个角色都被刻画得极端自私乃至近乎反社会人格。而一部在劳力士赞助的电影节上奢华首映、明星云集的

好莱坞影片发出这般警告，本身就令人五味杂陈。作为一部“信息电影”，《Artificial》恐怕只会印证AI末日论者最深的忧虑，同时激怒其坚定的拥护者。但影片确有几分真切的幽默，以刻薄的轻蔑审视着那些自封的未来工程师们——从他们笨拙的派对到他们怪异的步态。首映夜，巴林霍尔茨向观众送上了一句半心半意的讽刺安慰：“未来是光明的！我们都会没事的！”随即引来一阵极度紧张的笑声。

中文翻译：

《人工》是一部关于OpenAI和人工智能军备竞赛的新喜剧剧情片，是一部充满怪异步态的电影。

俄罗斯演员尤里·鲍里索夫饰演OpenAI联合创始人伊利亚·苏茨克维，他步履笨重地走来走去，像个怪咖版的弗兰肯斯坦怪物。安德鲁·加菲尔德饰演备受争议的OpenAI首席执行官山姆·奥特曼，他蹦蹦跳跳、孩子气十足——奥特曼本人最近还称赞过这种步态非常准确。而艾克·巴林霍尔茨饰演的埃隆·马斯克则迈着大步、趾高气扬，像某种灵长类首领一样指挥着一屋子电脑极客（他那长长的不满表情也与之相得益彰，让人想起那些老旧的“无聊猿”NFT）。

本周早些时候在纽约电影节首映后的问答环节中，巴林霍尔茨打趣地形容他的角色是“一个正常走路的人”。创造中的历史进程，就是这样尴尬地踱步前行。

《人工》由卢卡·瓜达尼诺执导，带着一些争议翩翩步入纽约电影节。这部电影原定由亚马逊米高梅发行，但在亚马逊与OpenAI达成500亿美元交易后，亚马逊在今年夏天引人注目地放弃了它。（独立制片公司Neon迅速接手。）纽约电影节艺术总监丹尼斯·林在台上介绍时提到了围绕这部电影的“那些情况”。

OpenAI可能对《人工》有意见，这可以理解。这部电影几乎完全不客气，将奥特曼和其他OpenAI利益相关者——以及所有推动AI热潮的主要人物——描绘成虚荣、鲁莽且无聊到令人难以忍受的人。2010年的《社交网络》成功让硅谷电脑天才看起来像是强大的行业巨擘和未来建筑师，而《人工》则朝着相反的方向用力。它将其科技巨头描绘成自恋的书呆子。

将其与大卫·芬奇的Facebook题材电影进行比较是不可避免的。当然，题材有相似之处。而且两部电影都由加菲尔德担任主角。（应该说，《人工》实际上更像是苏茨克维的传记片——他被描绘为当前AI浪潮的首要推动者，同时也是其喋喋不休的道德良知——而非奥特曼的传记片。）但瓜达尼诺和编剧兼制片人西蒙·里奇对素材的处理方式与芬奇截然相反。《社交网络》是一部冷峻、精致、台词密集的剧情片。《人工》则被定位为一部近乎超现实的黑色喜剧。

瓜达尼诺在好莱坞最知名的作品是成长剧情片《请以你的名字呼唤我》和三人行网球爱情片《挑战者》，他精通一种紧张的情欲表达。在《人工》中也有这种影子，体现在奥特曼对苏茨克维角色的诱惑性拉拢中。但这部电影在大多数情况下，无论从字面还是比喻意义上，都是无性的。

画面朴素，甚至有点丑陋。你几乎能感觉到瓜达尼诺更华丽的风格在蠢蠢欲动。一段慢镜头中，一屋子笨拙的程序员互相往身上浇Jarritos汽水，庆祝他们AI最新的自主学习里程碑——黏糊糊的，毫无性感可言。一场高潮戏中，奥特曼和他的OpenAI跟班们在旧金山一条斜坡街道上对峙苏茨克维，让人想起汤米·维索邪典经典《房间》的绿幕背景。放映结束后，瓜达尼诺声称这个故事捕捉的是另一种欲望——他称之为“取胜的欲望”。

这是一种许多观众可能共有的需求，因为AI行业的领导者们（包括奥特曼）似乎一致在压制这项技术的加速发展。我们被告知，人工智能发展如此迅速，变得如此精密，以至于有可能危及我们所知的地球生命。在最近接受《连线》杂志采访时，前Anthropic AI员工雅各布·考克森声称，他的前同事将未来几年描述为人类自身的潜在“终局”。另一位Anthropic研究员声称，AI有可能在十年内“杀死所有人类”。（概率被设定在低于10%，这仍然看起来极高。）

《人工》将这些紧张关系呈现为一种高尚的道德剧。鲍里索夫饰演的善意研究员——被描绘为性格上无法说谎——恳求设置保障措施、放慢速度，并让AI与人类更广泛的利益“对齐”。与此同时，支持和指数级技术与资本增长的理由，则由——基本上其他所有人来论证。

甚至安全本身这个理念也是可售卖的。在一个场景中，对OpenAI造成严重伤害能力的担忧被框定为只对公司利润有利。敲响警钟意味着透明度和尽职调查。而有害能力本身被视为技术进步的确实证据。这是双赢。同时也是双输。

《人工》就正在进行中的AI竞赛提出了一个令人信服且令人沮丧的观点。随便挑个比喻：火车已经离站，马已经跑出马厩，牙膏已经挤出管外。技术已经存在。因此，就像上世纪核武器一样，问题不再是力量本身，而是谁被允许驾驭它，以及如何驾驭。（在他离开Anthropic后的采访中，考克森直接将该公司比作曼哈顿计划。）

当所有参与者都按不同的规则行事时，伦理问题就变成了学术性的，或者只是令人厌烦的。如果不是OpenAI控制计算的未来，那就会是Anthropic，或者谷歌，或者埃隆·马斯克。或者，最糟糕的情况——至少根据特朗普政府的说法——中国。（《人工》的一个早期场景中，苏茨克维和他的导师、诺贝尔奖得主计算机科学家杰弗里·辛顿——由马克·里朗斯饰演——拒绝了一位渴望将他们的神经网络技术用于无人机和人脸识别系统的中国投资者的拉拢。）

对于普通观众来说——他们可能只接触AI来改写邮件或整理日程——很难判断《人工》的警示性警告到底该多认真对待。尤其是因为电影中基本上每个角色都被描绘为至少是完全自私自利的，如果不是反社会的话。当然，人们可能自然会对一部众星云集的好莱坞电影发出这些警告而感到有些不屑——它在一个由劳力士赞助的电影节上举行奢华首映，还有一场星光熠熠的派对，出席者包括明星、著名喜剧演员和知名电影人（包括八旬高龄的《出租车司机》编剧保罗·施拉德，他最近公开承认被一个“AI女友”甩了）。

作为一部传递信息的电影，《人工》似乎只可能证实AI末日论者最坏的担忧，同时激怒其铁杆支持者。然而，它在某些地方确实 genuinely 有趣，以一种轻蔑的态度看待我们未来大部分自封的工程师们——从他们傻气的派对到他们怪异的步态。

巴林霍尔茨也拥抱了这一立场，他在首映夜向观众献上了自己半心半意、充满讽刺的安慰：“未来是光明的！我们都会没事的！”随之而来的是一阵非常紧张的笑声。

评论

回到顶部

英文来源：

Artificial, the new comedy-drama about OpenAI and the artificial intelligence arms race, is a movie of weird walks.
As OpenAI cofounder Ilya Sutskever, Russian actor Yuri Borisov clomps around, like a geeky Frankensteinian monster. As the controversial OpenAI CEO Sam Altman, Andrew Garfield bounces and skips boyishly—a gait that Altman himself recently commended for its accuracy. And as Elon Musk, Ike Barinholtz puts on a loping swagger, commanding rooms of computer nerds like some sort of alpha gorilla (a semblance matched by his long, pouting faces, which call to mind those old Bored Ape NFTs).
During the movie’s post-premiere Q&A earlier this week at the New York Film Festival, Barinholtz wryly described his character as “a person who walks normally.” Such are the awkward saunters of the march of history in the making.
Artificial, directed by Luca Guadagnino, waltzed into NYFF with some controversy clinging to it. The film was originally set to be released by AmazonMGM, which conspicuously dropped it this summer after Amazon struck a $50 billion deal with OpenAI. (Indie house Neon swooped in to pick it up.) NYFF artistic director Dennis Lim referenced “the circumstances” around the film in his onstage introduction.
That OpenAI may have brooked issues with Artificial makes sense. The film is almost completely unflattering, depicting Altman and other OpenAI stakeholders—as well as all the major drivers of the AI boom—as vain, reckless, and insufferably boring. Where 2010’s The Social Network succeeded in making Silicon Valley computer savants look like robust titans of industry and architects of the future, Artificial works to opposing ends. It depicts its tech titans as self-regarding dweebs.
Comparisons to David Fincher’s Facebook drama are inevitable. There’s the subject matter, of course. And the fact that both films feature Garfield in a lead role. (It should be said that Artificial is really more of a biopic of Sutskever, here depicted as both the prime mover of the current AI wave and its nagging moral conscience, than it is of Altman.) But Guadagnino and writer-producer Simon Rich’s approach to the material feels diametrically opposed to Fincher’s. The Social Network was a steely, polished, wordy drama. Artificial is pitched as a black comedy that verges on the surreal.
Guadagnino, best known in Hollywood for the coming-of-age drama Call Me by Your Name and the three-way tennis romance Challengers, has mastered a strain of tense eroticism. There are shades of that in Artificial, in Altman’s seductive wooing of Sutskever’s character. But the film is, for the most part, literally and figuratively sexless.
The images are plain, and even a little ugly. You can almost feel Guadagnino’s more rococo style begging to break out. A slo-mo sequence of a room of gawky programmers dousing each other in bottles of Jarritos, celebrating their AI’s latest self-learning benchmark, is stickily unsexy. A climactic scene of Altman and his OpenAI lackeys confronting Sutskever on a sloping San Francisco street recalls the green-screen backdrops of Tommy Wiseau’s cult classic The Room. After the screening, Guadagnino claimed the story captured a different strain of desire—what he described as “the desire to prevail.”
It’s a need many viewers might share, as leaders in the AI industry (including Altman) appear united in tamping down the technology’s acceleration. Artificial intelligence, we are told, is developing so rapidly, and is becoming so sophisticated, that it risks compromising life on Earth as we know it. In a recent interview with WIRED, former Anthropic AI employee Jacob Coxon alleged that his former colleagues have described the next few years as the potential “endgame” for humanity itself. Another Anthropic researcher claimed that there’s a chance that AI could “kill all humans” within a decade. (The odds were set below 10 percent, which still seems extremely high.)
Artificial stages these tensions as a kind of a high moral drama. Borisov’s well-meaning researcher—depicted as being temperamentally incapable of lying—pleads for safeguards, slow-downs, and AI’s “alignment” with humanity’s broader interests. Meanwhile, the case for exponential technological and capital growth is argued by, well, basically everyone else.
Even the idea of safety itself as saleable. In one scene, concerns about OpenAI’s capacity for serious harm is framed as being beneficial only to the company’s bottom line. Sounding alarm bells signals transparency and due diligence. And the harmful capacity itself is seen as proof positive of the technology’s advancement. It’s win-win. And it’s also no-win.
Artificial makes a persuasive, and frustrating, point about the ongoing AI contest. Pick a metaphor: The train has already left the station, the horses have fled the barn, the toothpaste is out of the tube. The technology already exists. And so, as with nukes in the last century, it becomes not a question of the power itself but who is allowed to harness it, and how. (In his post-Anthropic exit interview, Coxon directly compared the company to the Manhattan Project.)
Questions of ethics become academic, or just annoying, when everyone involved is playing by a different rulebook. If it’s not OpenAI that controls computing’s future, it will be Anthropic, or Google, or Elon Musk. Or, worst of all—at least according to the Trump administration—China. (An early scene in Artificial sees Sutskever and his mentor, Nobel laureate computer scientist Geoffrey Hinton, played by Mark Rylance, rebuff the advances of a Chinese investor eager to use their neural network technology for drones and face-recognition systems.)
It may be hard for the average moviegoer, who may only engage with AI for email rewriting or calendar organization, to gauge just how seriously to take Artificial’s alarmist warnings. Not least of all because basically every character in the film is depicted as at least completely self-serving, if not sociopathic. And of course, one may naturally wince a little bit at a star-studded Hollywood movie raising these alarms, with a lavish premiere at a Rolex-sponsored film festival and a glitzy afterparty attended by stars, famous comedians, and notable filmmakers (including octogenarian Taxi Driver screenwriter Paul Schrader, who recently publicly confessed to being dumped by an “AI girlfriend”).
As a message movie, Artificial seems likely only to confirm the worst misgivings of AI doomers, while riling its diehard boosters. It is, however, genuinely funny in places, regarding the bulk of our future’s self-appointed engineers with withering contempt, from their dorky parties to their weird walks.
It was a stance embraced by Barinholtz, who offered the premiere-night audience his own half-hearted, sarcastic reassurance: “The future is bright! We’re all gonna be fine!” Cue a round of very nervous laughter.
Comments
Back to top