Explainables
Created with the help of AI
Category — Act

Feedback Loops

user agencyinteractive inquiry
User question

How can I correct the system?

Consulting signal

Comes up when a client wants to improve model quality over time using real user experience, or when they need to demonstrate to a regulator that errors are being captured and addressed.

Overview

Why this pattern exists

AI systems are not finished when they are deployed. They encounter new situations, make mistakes, and degrade over time as the world changes. The people best positioned to identify these problems are the ones using the system, and the people most affected by its mistakes are the ones subject to its decisions.

Feedback Loops is the pattern that connects users' experience of AI outputs back to the people responsible for the system. Not feedback as a compliance gesture ("we value your input"), but feedback as a genuine data pipeline: corrections, disagreements, and error reports that are collected, reviewed, and acted upon.

The design challenge is two-fold. First, the feedback mechanism must make it easy for users to signal when something is wrong: without requiring technical knowledge or significant effort. Second, it must be honest about what happens with feedback: closing the loop so users know their input was received and, where possible, what changed.

This pattern also addresses a subtler problem: feedback systems that collect input but never act on it erode trust faster than no feedback system at all. Users who submit corrections and receive no acknowledgment conclude that the system doesn't care. Design for the whole loop, not just the submission step.

Design goal

Provide a low-friction mechanism for users to report errors, provide corrections, and signal disagreement with AI outputs, and close the loop by communicating that feedback was received and, where appropriate, how it was used.

Usage guidance

When to use

  • The AI system makes predictions, classifications, or recommendations that may be factually incorrect or contextually wrong
  • Users have domain knowledge or direct experience that the model doesn't
  • The system is expected to improve over time and user feedback is a viable signal
  • Building user trust in the system is a design goal
  • The system operates in a context where users are experts (clinicians, analysts, domain professionals)

When not to use

  • The feedback cannot realistically be acted upon: only implement if there is a genuine process for reviewing and incorporating feedback
  • The decision is final and cannot be revised regardless of feedback
  • Feedback collection would expose users to privacy risks (e.g. their corrections reveal sensitive information)
  • The volume of feedback would exceed the organization's capacity to process it: design for the feedback volume you can actually handle

Design

UI primitives

Interactive Control / Rating

Thumbs up / down

The minimum viable feedback mechanism: a binary signal on a specific output. Appropriate for low-stakes or high-volume contexts where nuanced feedback is not feasible. Only use where the binary signal is genuinely informative.

Interactive Control / Button

Error flag / correction button

A specific action for flagging that an output is incorrect: "This is wrong" or "Flag an error." More informative than a binary rating because it signals a specific failure, not just dissatisfaction.

Interactive Control / Input Field

Correction field

An input field allowing users to provide the correct information alongside the error flag: "The AI said X. The correct answer is Y." Most valuable for systems where factual corrections are possible (entity recognition, data extraction, classification).

Interactive Control / Selector

Feedback reason selector

A structured dropdown of feedback categories: - Factually incorrect - Based on outdated information - Missing relevant context - Biased or unfair - Unclear or confusing - Other Structured reasons are more actionable for the review team than free-text alone.

Navigation & Flow / Confirmation

Feedback confirmation

After submission: a confirmation message that sets realistic expectations: "Your feedback has been received. We review all submissions weekly. This specific output may be updated in a future model release."

Contextual Overlay / Notification

What changed notification (optional)

When a user's feedback results in a system change, whether a correction or a model update, send a notification: "Based on feedback like yours, we updated [X] on [date]." Rare but powerful for trust-building.

Content Block / Panel

Aggregate feedback visibility

For transparency-forward systems: a public or accessible view of the volume and categories of feedback received, and what actions were taken. Demonstrates that the feedback loop is genuine.

How to use

Only implement feedback loops you can act on.

A feedback button that leads to a database no one reads is a trust-destroying pattern, not an explainability pattern. Design the feedback collection in conjunction with the review and update process.

Avoid participation washing.

Research in participatory AI identifies a failure mode where feedback collection functions as a performance of responsiveness rather than a genuine data pipeline: systems that ask for input to appear accountable without any mechanism for that input to change anything. Users who encounter this pattern, submitting feedback repeatedly with no visible effect, develop deeper distrust than users who were never offered a feedback mechanism at all. The design question is not "do we have a feedback button?" but "what is the path from this submission to a system change?"

Make the feedback mechanism specific to the output.

A generic "contact us" link is not a feedback loop. Feedback should be anchored to the specific output, decision, or claim it concerns.

Acknowledge receipt and set expectations.

Users who submit feedback and hear nothing assume it was ignored. A simple confirmation message with a realistic timeline ("reviewed monthly") is significantly better than silence.

Close the loop where possible.

Even a periodic email, for example "we've reviewed 200 pieces of feedback this month and updated the following categories," is more trustworthy than silence. Individual notifications are ideal where feasible.

Treat feedback as a data pipeline, not a complaint box.

Design the data structure of feedback collection so it can be queried, analyzed, and used to improve the model or process. Unstructured text feedback is hard to act on at scale.

Use cases

flow a

Clinician correcting a diagnostic suggestion

  1. 1. AI suggests diagnosis A.
  2. 2. Clinician disagrees based on additional context.
  3. 3. Clinician clicks "Flag concern": selects "Missing clinical context": adds a brief note.
  4. 4. Submits. Confirmation: "Your feedback has been logged and will be reviewed by the clinical AI team."
  5. 5. Case proceeds with the clinician's judgment, feedback noted in the record.
flow b

User correcting a data extraction error

  1. 1. System extracts information from a document and displays it.
  2. 2. User notices an extracted figure is wrong.
  3. 3. User clicks the correction icon next to the figure.
  4. 4. Correction field opens: user enters the correct value.
  5. 5. Correction is saved to the case, flagged for model review.
flow c

Content moderation feedback

  1. 1. User's content is flagged by an automated moderation system.
  2. 2. User believes this was a mistake.
  3. 3. Rather than a full appeal (see Contestability pattern 10), user submits a quick feedback: "This was flagged incorrectly: it does not violate [policy]."
  4. 4. Feedback is logged. If volume of similar flags exceeds a threshold, the policy rule is queued for review.

Design trade-offs

Participation vs. noise

Open feedback mechanisms can be flooded with low-quality or adversarial inputs. Design for signal quality: structured options, optional explanation, submission rate limits.

Transparency about use vs. managing expectations

Being honest about how feedback is used may reveal that individual feedback rarely leads to direct changes. Be truthful about timelines and probability of direct impact: users generally prefer honesty to false promises of responsiveness.

Individual correction vs. systemic improvement

Some feedback should fix this specific case immediately. Other feedback should improve the system for everyone. Design for both: immediate correction of individual errors where possible, and aggregated systemic improvement over time.

Connections

Relation to other patterns

Contestability challenges a specific adverse decision. Feedback is a lighter-weight signal about output quality. Both contribute to accountability but at different intensity levels and with different processes behind them.

Sources

establishes the theoretical and practical basis for human feedback as a mechanism for improving machine learning systems. Proposes that end users, not just data scientists, should be able to contribute corrections and signal errors. The foundational paper for this pattern. Published in AI Magazine, 35(4)

Birhane et al. (2022) — Participatory Machine Learning

argues that meaningful participation in AI systems requires more than feedback collection: it requires that affected communities have genuine influence over system design and governance. Motivates the "close the loop" design requirement

makes the case that accountability requires not just auditing AI systems but creating mechanisms through which failures can be reported and acted upon. Directly supports the feedback-as-accountability framing

Explainables
Created as a side project by Christian Laesser & AI