斯坦福大学伦理学中心主任莱夫·韦纳对有效利他主义（EA）运动提出了尖锐批评，指出其优良内核早已存在于古典慈悲与功利主义传统中，而其标新立异的部分却充斥着傲慢与逻辑漏洞。该运动常常沉溺于救助虾类等边缘议题或宏大的救世主幻想，在伪装成数学模型的复杂行话掩盖下，严重脱离现实常识。以知名评估机构GiveWell为例，该机构宣称其资助项目将挽救逾34万条生命，且今年拟发放高达10亿美元的资助，但其模型却排除了发展经济学熟知的负面溢出效应；例如其在尼日利亚北部资助的直接发放现金项目遭遇武装劫匪并导致2人死亡，这些伤亡在净收益计算中却未被核减为“负二”，且该机构的挽救生命结论从未经过同行评议审查。

该运动盲目追求效用最大化的教条在现实道德中遭遇了毁灭性崩塌。作为最彻底践行“利用一切资源实现效益最大化”信条的代表人物，加密货币从业者萨姆·班克曼-弗里德将他人视作可利用的工具，最终因巨额欺诈罪被判处25年联邦监禁。正如经济学家泰勒·考恩所质疑的，当有效利他主义领袖连身边的欺诈风险都毫无察觉时，社会根本无法信任其对人类长远未来的预测。班克曼-弗里德案的发生导致早期的EA领袖集体声誉扫地，使得这一流派在青年精英学生眼中变得尴尬而不可信。

韦纳强调，尽管该运动作为一个整体已趋于瓦解，但其成员掌控前沿人工智能核心权力的现实构成了严峻威胁。大量信徒涌入硅谷和Anthropic等头部AI实验室，Anthropic模型Claude的道德“宪法”起草者正是深涉EA思想的成员，而其近期安全问题的调查机构更与管理层的EA裙带圈子深度捆绑，缺乏独立外部制衡。韦纳呼吁年轻人不要沉迷于遥远的“长期主义”抽象推演，而应通过深入基层生活、参与地方政治或社区帮扶来积累真实生活经验；若缺乏现实判断力与制度约束，掌控AI的有效利他主义者恐将带来无法预估的灾难性后果。

Leif Wenar, head of Stanford University's ethics center, presents a trenchant critique of the effective altruism (EA) movement, arguing that its virtuous elements derive from ancient compassion and classical utilitarianism while its novel dogmas lack sound judgment. Characterized by saviour complexes and esoteric jargon, the movement prioritizes abstract causes over empirical grounding. For example, charity evaluator GiveWell claims its grants will save over 340,000 lives across poor nations and anticipates disbursing $1bn this year; however, its calculations omit systemic externalities and actual fatalities, such as failing to subtract two civilian deaths caused by armed bandits during a cash-transfer operation in northern Nigeria, while avoiding publication in peer-reviewed scientific journals.

The movement's moral and institutional credibility collapsed under its single-minded focus on resource exploitation. Sam Bankman-Fried, the crypto practitioner who epitomized EA's philosophy by viewing fellow human beings as instruments for capital generation to fund utilitarian causes, is currently serving a 25-year federal prison sentence for monumental fraud. As economist Tyler Cowen observed, a leadership cohort incapable of identifying immediate, severe fraud within their internal circles possesses zero credibility to project humanity's multi-millennial future. Consequently, the movement's intellectual leadership has been discredited, causing talented students across elite universities to increasingly distance themselves from its doctrines.

Wenar highlights the paramount risk that effective altruists now wield immense unilateral authority within frontier artificial intelligence laboratories. Many adherents have migrated into prominent tech companies like Anthropic, where the moral "constitution" governing the Claude system was authored by an individual immersed in EA dogma, while internal safety inquiries remain insulated within familial and ideological networks. Warning that the movement operates without sufficient external checks, Wenar urges students to cultivate practical judgment by living among real communities, engaging in local governance, or mentoring youth. He cautions that without accountability to lived reality, the technocratic hubris of EA adherents engineering AI could produce unprecedented human catastrophe.

Source: Can effective altruists grow up?

Subtitle: What AI needs is people of good judgment, writes Leif Wenar

Dateline: Oct 8th 2026\n