在面临一连串负面报导后，OpenAI 的 Sam Altman 正考虑暂缓最尖端人工智慧的研发，并向员工表示该公司可考虑放慢开发速度，且最好能与其他实验室同步进行。此举源于科技界对技术失控的持续警示，例如研究员 Jacob Coxon 指出此类技术在十年内有重大机率毁灭人类，而另一位 Anthropic 员工 Evan Hubinger 则评估该机率高于 10%（比率大于十分之一）。此外，Anthropic 发布的报告亦披露了敌对势力企图利用 Claude 模型优化网路攻击及制造武器的恶意行径。
反对暂缓发展的论点聚焦于激烈的市场竞争，若缺乏企业间的合作，单方面停滞可能导致美国实验室将优势拱手让给中国等竞争对手。同时，顶尖人工智慧开发商正积极推进公开上市计划，市场预期 Anthropic 本年度将寻求 2 兆美元（2,000,000,000,000 美元）估值挂牌，而 OpenAI 则计划于次年以超过 1 兆美元（1,000,000,000,000 美元）的估值走向公开市场，这使得任何研发上的停顿在商业层面上面临巨大的利益阻力。
金融界同样传出悲观论调，全球大型避险基金 Bridgewater 的共同投资长 Greg Jensen 示警，历史经验表明人类往往要等到人工智慧开始造成人员伤亡时才会采取防范措施。他强调面对这种必然的威胁，社会各界应在致命灾难发生前及早建立规范与应对机制，而非等到危机爆发后才被动回应。
Following a series of negative press reports, OpenAI's Sam Altman is considering pausing advanced artificial intelligence development, ideally coordinated with peer research laboratories. This consideration arises amid growing warnings from technology insiders regarding uncontrollable risks; researcher Jacob Coxon warned of potential human annihilation within this decade, while Anthropic's Evan Hubinger estimated the probability of existential catastrophe to exceed 10%. Concurrently, Anthropic released documentation detailing hostile attempts to leverage its Claude model for optimizing cyberattacks and weapon manufacturing.
Opposition to pausing development centers on intense market competition, where uncoordinated pauses by American entities could cede strategic technological advantages to Chinese competitors. Furthermore, primary developers face strong capital market incentives, as Anthropic targets an initial public offering this year seeking a valuation of $2 trillion, while OpenAI aims for a public listing next year valued at over $1 trillion, significantly complicating multilateral agreements to decelerate progress.
Echoing structural concerns from the financial sector, Greg Jensen, co-chief investment officer of Bridgewater, observed that historical precedent suggests regulatory interventions and defensive actions may not materialize until artificial intelligence directly causes human casualties. Jensen emphasized that proactive mitigation should precede fatal occurrences rather than relying on reactive measures after irreversible harm has already transpired.