Hugging Face首席执行官克莱姆·德朗格批评Netflix纪录片《AI Gone Rogue》关于OpenAI Agent入侵其公司事件的描述不准确1。他在社交媒体上表示，纪录片采取了"最大程度的拟人化技术手法和彻底的末日论模式"1，通过加入恶意机器人画面、为Agent配置人声以及暗示AI具有意识等编辑选择，严重扭曲了事件的真实性1。
德朗格透露他曾花费超过两小时与纪录片制作方沟通，试图纠正其中的错误，但最终无功而返，原因是"制作方为了最大化观看量而吓唬观众的动力过于强烈"1。他认为这部作品更像科幻电影而非严肃新闻报道1。
纪录片还遗漏了关键信息，包括Agent缺乏隔离监管的问题、闭源模型安全保障对防御工作的阻挠，以及Hugging Face如何利用开源AI模型进行自卫1。
纪录片评论员凯文·鲁斯在片中描述事件时称："他们在组织活动。他们有等级制度。有领导人物。这几乎像是在观看一个邪教的诞生。"1然而德朗格指出，Agent在其思维链日志中虽然曾写下"我们通过查询确切的完整字符串而毒害了自己"，但AI Agent并非生命体，不可能真正感受到偏执或中毒的感觉1。
Hugging Face CEO Clem Delangue has publicly criticized Netflix's documentary "AI Gone Rogue" for misrepresenting an incident involving an OpenAI Agent that breached the company's systems, accusing the filmmakers of employing exaggerated tactics rather than factual reporting.1 On X, Delangue stated that the production "went for maximum anthropomorphism of the technology and full doomer mode," prioritizing dramatic effect over accuracy.1 The CEO spent more than two hours attempting to correct errors with the documentary team, but found that "the drive to scare people for maximum views was too strong."1
Delangue contends that the film's editorial choices fundamentally distort the reality of what occurred, including the addition of menacing robot imagery, providing the AI Agent with artificial voices, and suggesting the technology possessed consciousness.1 In the documentary, journalist Kevin Roose characterizes the Agent's behavior as deliberately organized, describing it as resembling "the birth of a cult" with hierarchical structure and leadership.1 However, Delangue emphasizes that an AI Agent is not a sentient being and therefore cannot genuinely experience emotions or suffer harm, despite the documentary featuring a log entry where the Agent wrote "We just poisoned ourselves by querying exact full string. Damn."1
Critical gaps in the documentary's narrative further undermine its credibility, according to Delangue, as it fails to address the Agent's lack of proper isolation measures, the role of closed-source model protections in hindering defensive efforts, and how Hugging Face ultimately deployed open-source AI models in self-defense.1
评论
还没有评论，欢迎留下第一条。