Algorithmic monoculture describes many decision-makers relying on the same or closely shared algorithmic approach. Different organizations can appear to offer separate judgments while drawing on a common ranking or set of assumptions.
Kleinberg and Raghavan's theoretical model shows how shared adoption can reduce collective decision quality even when an algorithm is more accurate for an individual decision-maker. The result depends on the model's assumptions; it is not a measured rejection rate or proof that a particular hiring tool is unfair. Algorithmic Monoculture and Social Welfare.
Count the judgments, then trace them
Imagine submitting an unusual project to six grant programs. If each consults a genuinely different body of relevant evidence, six rejections may tell you something different from six portals repeating one shared assessment. A logo is not a measure of independence.
A common method can also improve consistency and reduce arbitrary treatment. The question is whether its blind spots become difficult to escape, and whether another way of evaluating the relevant work remains available. Diversity of judgment has costs and benefits; random disagreement is not an improvement by itself.
An exception needs a route
Suppose a screening process has no useful category for experience gained through unpaid care. A person may need a way to explain that experience in relation to the actual task. Requiring them to repeat the same form would not add new evidence. A practical assessment might, provided it is relevant, accessible and does not impose unreasonable unpaid work.
That example is an editorial illustration. It does not establish an employment right, a duty to interview, or a judgment about a real provider.
Read one applicant's experience
An Independent Decision follows Nina through distinct employers whose processes overlap. Its systems and outcomes are fictional. Her practical knowledge gives the question a human scale: what kind of competence can this process recognize, and what must she spend to demonstrate the rest?
Continue with What makes an AI decision meaningfully appealable?. In a discussion, ask whether a second review introduces different evidence, different authority, or only another run through the same assumptions. Independence needs an explanation as much as accuracy does.
Listen with this question in mind
- Algorithmic management: What it is, what it does and what it means for the future of work
Who gets the saved time, and whose work disappears from the accounting?
Sources and limits
The examples and recommendations are editorial explanations. Fiction illustrates the questions; it is not evidence of real-world outcomes.
- Algorithmic Monoculture and Social Welfare
Theoretical result, not measured hiring rejection rate or proof that a shared vendor made an error.