Building the AI Scientist for the Hardest Part of Science

The industry is converging on the idea of an AI scientist, even if the label and scope are not yet consistent. Google’s AI co-scientist is a Gemini-based multi-agent system for generating and refining hypotheses and research proposals. Edison Scientific calls Kosmos an AI Scientist for autonomous, data-driven discovery. Benchling describes The AI Scientist as wiring together the digital and physical worlds of R&D. Anthropic calls Claude Science an AI workbench for scientists that integrates research tools, produces auditable artifacts, and manages access to computing resources. These systems use different language and emphasize different capabilities, which is why the market needs a clearer definition of what an AI scientist should be able to do.

The best way to understand the fundamental capabilities of an AI scientist is to compare two scientific questions:

1. 01

   Given an amino-acid sequence, what three-dimensional structure will the protein adopt?

2. 02

   Why do some patients with fibrosis respond to a drug while others do not, and what should the development team do next?

Both are difficult scientific questions, but their problem structures are different. The protein question resembles navigation across an enormous terrain toward a defined destination. The input and desired output are specified, candidate structures can be evaluated against learned geometric and biological constraints, and the system can narrow the space of possibilities. It is primarily a search, inference and optimization problem with a defined output.

The fibrosis question resembles detective work. The relevant clues are distributed across experimental data, patient biology, prior literature and clinical context. Some clues are missing, several explanations may fit what is known, and the scientist must decide which pieces belong together. The problem is like a jigsaw puzzle, but with more than one plausible arrangement. The work is to construct and challenge testable explanations from incomplete context, then identify the new observation that would distinguish them.

Protein-structure prediction is one of the most consequential achievements in modern computational biology. AlphaFold frames structure prediction as graph inference in three-dimensional space, using learned representations of sequence and residue relationships to predict a structural model. Modern systems do not enumerate every possible conformation. They reduce the effective search space and produce a candidate structure that can be assessed through confidence estimates, benchmark performance and comparison with experimentally determined structures.

The fibrosis question has a different structure. Fibrosis can arise from different injuries and etiologies, and the active mechanism and the relevance of a model depend on biological context. A development team observing that some patients respond while others do not must consider disease stage, active mechanism, tissue context, prior treatment, drug exposure, comorbid biology and the measurements used to define response. A molecular signature could identify causal biology, reflect a downstream consequence, track drug exposure, or mark a correlated patient characteristic. The evidence does not arrive with labels that tell the scientist which interpretation is correct.

Humans call the next step forming a hypothesis. That chain is not proof. It is a proposition with assumptions and testable predictions that must be compared with other explanations and tied to an observation that could prove it wrong.

Much of the current market discussion emphasizes the computational side more strongly than the other. Systems may be excellent at running tools, writing code, searching large spaces, or optimizing toward a specified output without yet being equally strong at deciding which biological context matters and constructing a causal explanation from incomplete evidence. Coding capability remains an important primitive, but it is not the whole architecture.

The work product for the fibrosis team should be a reviewable scientific position: which mechanism best explains differential response, which patient subgroup the explanation applies to, what supports and contradicts it, which alternatives remain credible, and what evidence would change the recommendation. This is what takes a system beyond computation without diminishing the importance of computation itself.

A response-associated signal can support several explanations. Responders may have target-dependent biology that the drug successfully modifies. Non-responders may have inadequate exposure despite sharing the same mechanism. A parallel immune process may sustain disease even when the intended target is engaged. Each explanation can fit part of the observation and lead to a different next step.

An evidence matrix helps piece those clues together and makes the reasoning inspectable. Its rows are observations and its columns are competing hypotheses. Each observation can support a hypothesis, weaken it, remain compatible with it, or leave the relationship uncertain. The purpose is to prevent the system from presenting only the evidence that fits its preferred answer, expose uncertainty, and show where the explanations make different predictions.

The discriminating experiment follows from those differences. If target engagement is confirmed in both responders and non-responders, inadequate exposure becomes less plausible. If non-responders still show an immune activation program, a parallel mechanism becomes more plausible. The next experiment is selected because its possible outcomes would update the ranking of the hypotheses.

A long-running agent or a strong foundation model is only one component of an AI scientist. The wider architecture requires:

• computation
• scientific retrieval
• knowledge graphs
• provenance
• episodic and semantic memory
• context compression
• causal reasoning
• evaluation
• scientific discourse
• governed execution

These capabilities allow the system to carry the path from a reproducible result to a challengeable scientific interpretation.

A system that performs the first task has valuable computational capability. A system that can also carry the second task through evidence, causal reasoning, falsification and human review is beginning to operate as an AI scientist. That distinction should define the category and the standard by which the market evaluates it.