近期人工智慧专家纷纷警告AI所带来的生存风险，特别是生物武器威胁已超越传统硬体武器的想像，调查显示高达七成的国安专家认为AI正显著提升生物威胁的风险。以往制造危险病原体高度依赖难以透过书本获取的「隐性知识」与实验室学徒制，但AI基础模型与自动化实验室的快速发展，正迅速消除这些技术门槛，使攻击者更有可能轻易合成出致命病毒。（关键数字：70）

为因应此项严峻危机，专家呼吁应强制实施全面的合成DNA筛检机制，防止病原体基因序列被恶意制造，并建立漏洞悬赏制度与自动化实验室验证规范。然而，目前美国政府缺乏完善的监管作为，联邦审查架构长年搁置，甚至缩编了相关的生物防护组织，导致关键的安全政策无法制度化推进。

在官方监管缺位的当下，民间与私营企业必须主动填补安全缺口以自律防范未然。例如前线AI机构已主动拦截多起利用模型辅助生物武器研发的尝试，科技领袖必须在更严格的政府干预或灾难性事故爆发之前，及早落实各项防护屏障，避免致命生物灾害的发生。

Recent warnings from artificial intelligence experts highlight that existential AI risks lie primarily in biological threats rather than robotic warfare, with 70% of national security practitioners agreeing that AI meaningfully elevates bioweapon risks. Historically, synthesizing lethal pathogens was constrained by specialized 'tacit knowledge' that demanded direct lab experience, but modern foundation models and automated laboratories are dismantling these barriers and lowering technical hurdles for malicious actors.

To counter this escalating danger, experts advocate for mandatory DNA synthesis screening, bug bounties to test screening systems, and strict verification protocols for automated laboratories. Nevertheless, federal governance has stalled revised screening frameworks, dismantled core biosecurity bodies, and resisted enacting critical regulatory safeguards, creating a dangerous policy vacuum.

Given the absence of prompt federal oversight, private industry leaders must step forward to implement self-regulatory safety guardrails. Frontier AI firms have already begun intercepting attempts to exploit language models for bioweapon research, and companies must continue acting proactively before catastrophic biological incidents occur or crippling regulations are imposed.