科学领域的 AI 需要推理，而不只是数据

诚然，AI 将为科学带来非同寻常的变化，但越来越清楚的是，AlphaFold 及类似系统或许并不是这场变革的最佳范式。尽管 AlphaFold 是一项深刻的成就，但促成它诞生的条件十分罕见；在其他领域满足这些条件所需的时间将以数十年计，而不是数年。科学加速更可能来自另一种路径：AI Agent。

AlphaFold 成功的首要条件，是 Protein Data Bank 的存在——这是一个包含约 17 万个经实验验证的蛋白质结构的数据集，DeepMind 团队据此训练了模型。Protein Data Bank 的创建绝非易事：它经历了 53 年的国际科学合作，并且根据一项近期估算，汇集这些数据需要价值约 210 亿美元的实验工作。这种规模的工作出了名地难以获得资金支持，几乎不可能协调，并且实施起来极为耗时；因此，它们往往以失败告终。

但即便是在具备必要凝聚力与资源、且相关数据没有因商业所有权而无法获取的领域，另一个障碍也很少被讨论：在科学上根本无法生成可比的数据。对于蛋白质结构而言，关键实验技术——蛋白质晶体学——是一种异常可重复且可靠的工具，以至于已有超过 25 项诺贝尔奖依赖于它。但在大多数实验科学中，结果往往并不稳定。细胞系会发生漂移，化学品会含有痕量污染物，实验室湿度会变化。要在生物学或大多数化学领域创建足够一致、足够准确、足够精密且足够可扩展的测量数据集，以训练现代神经网络，将需要新型测量手段和新的标准化方法——而这些短期内都不会准备就绪。

当然，确实有少数领域满足这些要求：天气预报、大部分基因组学，以及极少数化学领域。如果它们尚未实现，就可能很快迎来 AlphaFold 式的突破。正如美国国家新兴生物技术安全委员会所主张，政府对这些数据集生产和协调工作的支持将至关重要。但对于科学中的大多数开放问题，至少在短期内，我们需要另一套方案。幸运的是，一种更低调、更务实的方法已经开始显现前景。

科学家一直都在不确定性下进行推理。试图识别新药物靶点的生物学家从未拥有完美的数据集。相反，他们会结合对接计算和已知结构，纳入分子动力学因素，进行少量结合测定，并依据每种方法特有的优势和失效点运用判断力进行权衡。科学的技能不在于任何单一工具，而在于综合多种工具产出的结果，并随着证据出现而修正结论。这正是绝大多数实际科研工作的推进方式。但直到最近，还没有任何软件能够做到这一点。

如今，Agent 可以做到。简单来说，Agent 是一种被赋予工具访问权限——无论是数字工具还是物理工具——以及使用这些工具能力的 AI 推理引擎。在过去几年中，AI 的一项基础性架构转变推动了这类程序的快速涌现。它们由大语言模型驱动，大幅降低了对科学领域专用数据集的需求。对科学而言，这项技术进步代表着一项基础性变化：它让我们能够创建数字工具，以模拟真实科研中迭代推进、高度依赖具体情境的过程。AlphaFold 等工具将强大的方法应用于有限的问题，而 Agent 天生是通才。它们并不代表一种全新的科学研究方式；相反，它们以数字形式建模了人类发现的过程。

以 Google 在 5 月宣布的 AI Co-Scientist 为例。研究人员向其提供了一份一页纸的简报和一个目标：弄清抗生素耐药性如何在细菌物种之间传播——这是耐药性感染的关键驱动因素。该系统启动了多个子 Agent：一个根据文献起草假设；另一个像同行评审者一样挑错；第三个通过竞赛对最有力的候选假设进行排序；第四个则完善胜出的假设。该 Agent 得出结论：耐药基因正在借助细菌病毒“搭便车”，利用任何能够将其运送到新宿主的病毒。这个假设是正确的。伦敦帝国理工学院的研究人员曾通过艰苦的湿实验室工作，用十年时间得出同样结论；而他们的论文此前未被 Co-Scientist 看到，当时仍在同行评审中。

像 Co-Scientist 这样的 Agent 仍是新型工具。在它们成为科学流程中无处不在的一部分之前，仍有切实的挑战需要克服：它们依然可能产生幻觉，判断并不稳定，其记忆和输入限制也约束了它们能够自主运行的时间。但这些技术障碍终将消退；随着它们消退，我们将开始看到科学 Agent 对科研可靠性、一致性和速度所带来的复合效应。

To be sure, AI will bring extraordinary changes to science, but it has become increasingly clear that AlphaFold, and things like it, may not be the best template for that metamorphosis. Though it is a profound achievement, the conditions that produced the likes of AlphaFold are rare, and the time it will take to meet those conditions in other fields will be measured in decades, not years. Instead, the acceleration of science will come about thanks to another approach: AI agents.

The primary condition for AlphaFold’s success was the existence of the Protein Data Bank, a data set of roughly 170,000 experimentally validated protein structures on which DeepMind’s team could train its model. The creation of the Protein Data Bank was not simple: It took 53 years of international scientific cooperation and, by a recent estimate, roughly $21 billion worth of experimental work to assemble. Efforts of that scale are infamously difficult to fund, next to impossible to coordinate, and hugely time-consuming to execute; they have often been unsuccessful as a result.

But even in fields with the requisite cohesion and resources, and where the relevant data are not rendered inaccessible by commercial ownership, another barrier is too little discussed: the scientific impossibility of generating comparable data. In the case of protein structures, the key experimental technique—protein crystallography—is an unusually replicable and dependable tool, so much so that over 25 Nobel Prizes have relied on it. But in most of experimental science, results vary more often than not. Cell lines drift. Chemicals have trace contaminants. Lab humidity changes. The creation of measured datasets that will be consistent enough, accurate enough, precise enough, and scalable enough to train a modern neural network in biology or most of chemistry would require new kinds of measurement and new standardized approaches—none of which will be ready anytime soon.

Of course, there are a handful of fields where these requirements are met: weather forecasting, much of genomics, very limited areas of chemistry. These may see AlphaFold-style breakthroughs soon, if they haven’t already. Government support for the production and coordination of those datasets will be critical, as the US National Security Commission on Emerging Biotechnology has argued. But for most open questions in science, we will need a different plan, at least in the short term. Luckily, something quieter and more modest has begun to show promise.

Scientists have always reasoned under uncertainty. Biologists working to identify new drug targets have never had perfect datasets. Instead, they combine docking calculations and known structures, factor in molecular dynamics, run a handful of binding assays, and use their judgment to weigh each method according to its particular strengths and points of failure. The skill of science is not in any single tool; it is synthesizing what many tools produce, and revising the results as the evidence comes in. This is how most working research actually proceeds. But until very recently, no software could do it.

Agents now can. Simply put, an agent is an AI reasoning engine that has been given access to tools—digital or physical—and the capabilities to use them. Over the last few years, a fundamental architectural shift in AI has enabled the rapid proliferation of these programs, which are powered by large language models, dramatically reducing the need for scientifically specialized datasets. For science, this technological advancement represents a foundational change: it has allowed us to create digital tools that can mimic the iterative, highly contingent process of actual research. While tools like AlphaFold apply a powerful approach to a limited question, agents are inherently generalists. They do not represent a new way to do science—instead, they digitally model the human process of discovery.

Consider Google’s AI Co-Scientist, announced in May. Researchers gave it a one-page brief and a goal: Figure out how antibiotic resistance spreads between bacterial species, a key driver of drug-resistant infections. The system spun up sub-agents. One drafted hypotheses from the literature. Another picked them apart like a peer reviewer. A third ran tournaments to rank the strongest candidates. A fourth refined the winning hypothesis. The agent concluded that resistance genes were hitching rides on bacterial viruses, borrowing whichever virus could ferry them into a new host. The hypothesis was correct. Researchers at Imperial College London had spent a decade reaching the same conclusion through painstaking wet-lab work; their paper, previously unseen by Co-Scientist, was still in peer review.

Agents like Co-Scientist are still novel tools, and there are real challenges to overcome before they become a ubiquitous part of the scientific process: They are still liable to hallucinate, their judgment is not consistent, and they have memory and input constraints that limit the time they can run autonomously. But these technical barriers will fall away, and as they do we will begin to notice the compounding effects of scientific agents on the reliability, consistency, and velocity with which science is done.