How technical choices can lead to safety failures, and what to do differently The post AI Harms and Their Technical Causes appeared first on Towards Data Science.
How technical choices can lead to safety failures, and what to do differently The post AI Harms and Their Technical Causes appeared first on Towards Data Science.
In context
- Topic: Datos y Analítica — Gobierno del dato, calidad, analítica y productos de datos.
- Source: Towards Data Science
- Published: 19/10/2026
Continue reading at the original source →
Excerpt published automatically by the site radar. The full text belongs to its publisher and is linked above.
Why it matters
A beautiful dashboard can hide a bad number for months. That is why I distrust stories about visualisation that say nothing about governance: the expensive part is not displaying the number, it is guaranteeing the number means the same thing on Tuesday as on Thursday.
When I assess a data initiative, the first question is which concrete decision gets made with this and who makes it. If the answer is that it provides visibility, I already know the dashboard will be abandoned within a year. Visibility is not an objective, it is a side effect.
What usually goes wrong
Where it usually breaks is capture. A lot gets invested in processing and nothing at the point where a person fills in the form, in a hurry, with no validation and no idea what it is for. No model fixes data that was born wrong; it only propagates it with more authority.
What to watch
- Where the data comes from and what happens when the source system changes or fails.
- When the number arrives relative to when the decision is taken, which is what defines its usefulness.
- Who answers for the quality of the indicator when somebody challenges it in a committee.
How I read this entry
I would use it to review the full cycle, from where the data is captured to where somebody acts. The problem is almost never in the analytics but at the start: someone filling a field by hand, with no validation, in a hurry, not knowing what it is used for later.
This entry is an excerpt from the original source, selected by the site radar. The commentary above is the site's own and does not belong to the cited publisher.
Comments
No comments yet. Yours can be the first.
Living through this in your own team?
Open the chat and tell me how you're handling it. I'm interested in comparing notes.
