Keep human judgment at the center of science and engineering.
AI can now produce code, models, simulations, analyses, and explanations quickly. But producing an output is not the same as understanding the problem, evaluating the method, or knowing whether the result is sound.
Our goal is to help scientists, engineers, educators, and students use AI as part of their work without removing the judgment and responsibility that technical problem-solving requires. WEEMS is being developed around a simple principle: AI should make difficult work more manageable, while people remain responsible for what is being studied, how the work is carried out, and what conclusions should be drawn.
What we are building
WEEMS brings computational work into one project: notebooks, data, compute sessions, scientific software, the models and agents used in the work, and results.
We are also exploring a broader direction we call judgment-centered computing. Scientific and engineering results depend on decisions about problem formulation, assumptions, methods, validation, and interpretation. These decisions should not disappear inside code, prompts, or automated workflows.
As models and agents take on more of the work, the choices shaping that work can become less visible. A model may select an approach, introduce assumptions, or interpret an output without making clear which parts were computed and which required judgment. WEEMS is being designed so that models and agents can investigate alternatives, carry out computations, and show how different choices affect the result, while consequential technical decisions remain visible, revisable, and under the user’s direction.