[论文解读] Artificial intelligence and biological misuse: Differentiating risks of language models and biological design tools

Jonas B. Sandbrink|arXiv (Cornell University)|Jun 24, 2023

Law, AI, and Intellectual Property被引用 22

一句话总结

本文区分两类 AI 工具的风险——大型语言模型（LLMs）和生物设计工具（BDTs）——并讨论它们的汇聚如何提高生物安全风险，以及提出的干预措施。

ABSTRACT

As advancements in artificial intelligence (AI) propel progress in the life sciences, they may also enable the weaponisation and misuse of biological agents. This article differentiates two classes of AI tools that could pose such biosecurity risks: large language models (LLMs) and biological design tools (BDTs). LLMs, such as GPT-4 and its successors, might provide dual-use information and thus remove some barriers encountered by historical biological weapons efforts. As LLMs are turned into multi-modal lab assistants and autonomous science tools, this will increase their ability to support non-experts in performing laboratory work. Thus, LLMs may in particular lower barriers to biological misuse. In contrast, BDTs will expand the capabilities of sophisticated actors. Concretely, BDTs may enable the creation of pandemic pathogens substantially worse than anything seen to date and could enable forms of more predictable and targeted biological weapons. In combination, the convergence of LLMs and BDTs could raise the ceiling of harm from biological agents and could make them broadly accessible. A range of interventions would help to manage risks. Independent pre-release evaluations could help understand the capabilities of models and the effectiveness of safeguards. Options for differentiated access to such tools should be carefully weighed with the benefits of openly releasing systems. Lastly, essential for mitigating risks will be universal and enhanced screening of gene synthesis products.

研究动机与目标

• 区分 LLMs 相对于生物设计工具（BDTs）所带来的生物安全风险。
• 解释将 LLM 作为实验室助手或自主科学工具时，如何降低被误用的门槛。
• 描述 BDTs 如何使得更危险的病原体和定向武器的创建成为可能。
• 提出管理风险的干预措施，包括发行前评估和对基因合成产品的筛查。

提出的方法

• 基于能力和潜在的滥用情景比较 LLMs 与 BDTs 的风险概况。
• 主张进行独立的发行前评估以评估模型能力和防护措施的有效性。
• 讨论包括对工具的差异化访问和开放性考量的政策选项。
• 倡导对基因合成产品的普遍且加强筛查以减轻风险。
• 在将 LLMs 与 BDTs 共同使用时分析潜在的汇聚效应。

实验结果

研究问题

• RQ1 与大型语言模型相比，生物设计工具相关的生物安全风险有哪些区别？
• RQ2 LLMs 与 BDTs 的汇聚如何增加对生物因子的潜在伤害？
• RQ3 在不妨碍有益 AI 进展的前提下，哪些干预措施可以有效降低风险？
• RQ4 对用于生物学领域的强大 AI 工具的访问，包括发行前评估和筛查程序，应该有哪些适当的治理选项？

主要发现

• LLMs 可能提供双重用途信息，并在被用作实验室助手或自主科学工具时降低生物误用的门槛。
• BDTs 可能使得比迄今为止看到的更糟的流行病原体得以创建，并支持更可预测、定向的生物武器。
• LLMs 与 BDTs 的汇聚可能抬高伤害上限并扩大危险能力的可及性。
• 独立的发行前评估可以帮助评估模型能力与防护措施；差异化的访问政策应该权衡开放发布的益处。
• 增强对基因合成产品的筛查对于降低风险至关重要。