«

逝者虽已免受诽谤,但这绝非伪造其音容的正当理由。

qimuai 发布于 阅读:9 一手编译


逝者虽已免受诽谤,但这绝非伪造其音容的正当理由。

内容来源:https://techcrunch.com/2025/10/07/you-cant-libel-the-dead-but-that-doesnt-mean-you-should-deepfake-them/

内容总结:

已故喜剧大师罗宾·威廉姆斯的女儿泽尔达·威廉姆斯近日通过社交媒体发出痛心呼吁,请求公众停止向其发送利用人工智能技术生成的父亲影像。她在Instagram动态中直言:“这些视频不仅令我感到不适,更违背了父亲的意愿。若尚存一丝尊重,请立即停止这种行为。”

这一发声恰逢OpenAI发布新一代视频模型Sora 2及同名社交应用之际。该技术允许用户生成高度逼真的深度伪造视频,虽然对在世名人设有限制,却为已故公众人物的数字复活开启方便之门。平台测试显示,马丁·路德·金、约翰·列侬等历史人物及罗宾·威廉姆斯等已逝明星的形象均被肆意滥用。

值得关注的是,尽管OpenAI为在世者设置了“数字形象授权”防护机制,逝者却无法享有同等保护。更令人担忧的是,根据美国学生新闻法制中心说明,法律上对逝者名誉权的保护存在空白,这使得相关行为难以追责。

泽尔达对此痛心疾首:“目睹真实人物的遗产被简化为‘形似声近即可’的电子玩偶,任由他人操纵制作低劣内容,实在令人愤怒。”她的遭遇折射出当前AI技术发展中的伦理困境。

面对业内批评,OpenAI最初声称影视公司需主动申请才能排除其IP被滥用,后在电影协会援引著作权法施压后转变立场。但Sora模型生成的视频已达到以假乱真程度,其潜在危害远超其他技术平台。有专家警告,若放任技术将生者与逝者皆视为数字玩物,或将开创危险的先例。

(注:原文中关于科技大会的广告内容及部分重复表述已根据新闻报道规范予以删减整合。)

中文翻译:

已故演员罗宾·威廉姆斯的女儿塞尔达·威廉姆斯近日向父亲影迷发出痛心呼吁。她在社交媒体Ins上写道:"求你们别再给我发用AI技术合成的父亲视频了。别以为我想看或能理解这种行为——我不想看也无法理解。如果你们还有起码的良知,请停止对他、对我、对所有人做这种事。这既愚蠢又浪费精力,而且相信我,这绝不是他愿意看到的。"

她选择在此时发声并非偶然——几天前OpenAI刚发布Sora 2视频模型及同名社交应用,该技术能让用户生成自己、朋友及特定卡通形象的高度逼真深度伪造视频。据美国学生新闻法律中心解释,逝者成为伪造目标的现象尤为突出,因为法律上并不认定诽谤逝者构成违法。

Sora禁止生成在世者视频(除非获得本人或友人授权),但对逝者几乎毫无限制。目前仍处内测阶段的应用中,已涌现大量历史人物与已故名人的伪造视频,从马丁·路德·金、罗斯福到约翰·列侬、罗宾·威廉姆斯等。

TechCrunch测试发现,Sora 2虽拒绝生成2024年去世的前总统吉米·卡特等人物视频,却允许生成2014年逝世的罗宾·威廉姆斯形象。尽管OpenAI设置了"数字形象授权"功能供在世者控制自身影像使用,逝者却无法享有同等保护。若尼克松目睹笔者伪造他主张废除警方的视频,恐怕在坟墓里都要辗转反侧。

OpenAI对逝者深度伪造的合规标准令人困惑。尽管该公司未回应TechCrunch的质询,但法律先例表明其很可能无需为诽谤逝者承担责任。当科技公司竞相追赶OpenAI时,若我们将真人——无论生死——视为可随意摆弄的玩偶,必将开创危险的先例。

"目睹真实人物的毕生成就被简化为'长得像声音像就够用了',好让某些人在TikTok上炮制操纵他们的拙劣内容,这令人发狂。"塞尔达写道。批评者指责OpenAI在版权问题上采取放任态度,致使Sora上线即涌现大量彼得·格里芬、皮卡丘等版权角色的AI视频。虽然CEO阿尔特曼最初要求好莱坞工作室需主动选择退出IP保护,后在电影协会严正声明下转变立场。

作为当前最危险的深度伪造工具,Sora生成内容的逼真度远超xAI等竞品。当其他平台以更宽松的监管追赶时,若放任真人影像沦为数字玩物,人类将滑向伦理的深渊。

英文来源:

Zelda Williams, daughter of the late actor Robin Williams, has a poignant message for her father’s fans.
“Please, just stop sending me AI videos of Dad. Stop believing I wanna see it or that I’ll understand. I don’t and I won’t,” she wrote in a post on her Instagram story on Monday. “If you’ve got any decency, just stop doing this to him and to me, to everyone even, full stop. It’s dumb, it’s a waste of time and energy, and believe me, it’s NOT what he’d want.”
It’s probably not a coincidence that Williams was moved to post this just days after the release of OpenAI’s Sora 2 video model and Sora social app, which gives users the power to generate highly realistic deepfakes of themselves, their friends, and certain cartoon characters.
That also includes dead people, who are seemingly fair game because it is not illegal to libel the deceased, according to the Student Press Law Center.
Sora will not let you generate videos of living people — unless it is of yourself, or a friend who has given you permission to use their likeness (or “cameo,” as OpenAI calls it). But these limits don’t apply to the dead, who can mostly be generated without roadblocks. The app, which is still only available via invite, has been flooded with videos of historical figures like Martin Luther King, Jr., Franklin Delano Roosevelt, and Richard Nixon, as well as deceased celebrities like Bob Ross, John Lennon, Alex Trebek, and yes, Robin Williams.
How OpenAI draws the line on generating videos of the dead is unclear. Sora 2 won’t, for example, generate former President Jimmy Carter, who died in 2024, or Michael Jackson, who died in 2009, though it did create videos with the likeness of Robin Williams, who died in 2014, according to TechCrunch’s tests. And while OpenAI’s cameo feature allows people to set instructions for how they appear in videos others generate of them — guardrails that came in response to early criticism of Sora — the deceased have no such say. I’ll bet Richard Nixon would be rolling over in his grave if he could see the deepfake I made of him advocating for police abolition.
OpenAI did not respond to TechCrunch’s request for comment on the permissibility of deepfaking dead people. However, it’s possible that deepfaking dead celebrities like Williams is within the firm’s acceptable practices; legal precedent shows that the company likely wouldn’t be held liable for the defamation of the deceased.
Join 10k+ tech and VC leaders for growth and connections at Disrupt 2025
Netflix, Box, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla — just some of the 250+ heavy hitters leading 200+ sessions designed to deliver the insights that fuel startup growth and sharpen your edge. Don’t miss the 20th anniversary of TechCrunch, and a chance to learn from the top voices in tech. Grab your ticket before doors open to save up to $444.
Join 10k+ tech and VC leaders for growth and connections at Disrupt 2025
Netflix, Box, a16z, ElevenLabs, Wayve, Hugging Face, Elad Gil, Vinod Khosla — just some of the 250+ heavy hitters leading 200+ sessions designed to deliver the insights that fuel startup growth and sharpen your edge. Don’t miss a chance to learn from the top voices in tech. Grab your ticket before doors open to save up to $444.
“To watch the legacies of real people be condensed down to ‘this vaguely looks and sounds like them so that’s enough,’ just so other people can churn out horrible TikTok slop puppeteering them is maddening,” Williams wrote.
OpenAI’s critics accuse the company of taking a fast-and-loose approach on such issues, which is why Sora was quickly flooded with AI clips of copyrighted characters like Peter Griffin and Pikachu upon its release. CEO Sam Altman originally said that Hollywood studios and agencies would need to explicitly opt out if they didn’t want their IP to be included in Sora-generated videos. The Motion Picture Association has already called on OpenAI to take action on this issue, declaring in a statement that “well-established copyright law safeguards the rights of creators and applies here.” He has since said the company will reverse this position.
Sora is, perhaps, the most dangerous deepfake-capable AI model accessible to people so far, given how realistic its outputs are. Other platforms like xAI lag behind, but have even fewer guardrails than Sora, making it possible to generate pornographic deepfakes of real people. As other companies catch up to OpenAI, we will set a horrifying precedent if we treat real people — living or dead — like our own personal playthings.

TechCrunchAI大撞车

文章目录


    扫描二维码,在手机上阅读