Difference between revisions of "Attribute"
From MgmtWiki
(→Solutions) |
(→Problems) |
||
| Line 13: | Line 13: | ||
* Any attribute about a digital entity can be used to narrow the population that exhibits that attribute. | * Any attribute about a digital entity can be used to narrow the population that exhibits that attribute. | ||
* If you want to see how little data is needed to uniquely determine your real world identity, or your preferences, just enter your data into [https://aboutmyinfo.org/index.html this little tool]]. | * If you want to see how little data is needed to uniquely determine your real world identity, or your preferences, just enter your data into [https://aboutmyinfo.org/index.html this little tool]]. | ||
| + | |||
| + | Every attribute is a generalization, a categorization, a stereotype, a limitation on the real person to whom it applies. Human beings all have evolved to form generalizations about our world, it is how we make sense of the world. It is how we learn about the world. It is how we talk about the world. But, in the final analysis, all generalizations are at least partially wrong. Each person has their real identity and as we learn about them, we find new things about them. To remember we assign that person additional attributes so that we can remember them and so that we can talk about them. However we model attributes in the digital world, it needs to conform to the use that humans make of attributes in the real world. | ||
==Solutions== | ==Solutions== | ||
Revision as of 11:18, 2 February 2019
Full Title or Meme
Any datum (piece of data) about a digital entity.
Context
- At one time Attributes were once considered to be a useful way to perform Authentication of a User.[1]
- Now it is realized that this method releases much User Private Information and offers low Assurance.
- Even the ISO definition of Personally Identifiable Information (or PII) seems to imply the existence of Attributes that are not personally identifying.
- There are two main sources of user Attributes:
- The user is often asked to supply Attributes, sometimes with a high level of Assurance. The Attribute may come directly from the user or from an Identifier or Attribute Provider. In either case User Consent is easy to request and control.
- External sources generate User Information which could become Attributes that help identify them, for example EHI or electronic health information from lab tests or health provider notes. User Consent is harder to track and control since the user may not even understand the data generated or its implications.
Problems
- Any attribute about a digital entity can be used to narrow the population that exhibits that attribute.
- If you want to see how little data is needed to uniquely determine your real world identity, or your preferences, just enter your data into this little tool].
Every attribute is a generalization, a categorization, a stereotype, a limitation on the real person to whom it applies. Human beings all have evolved to form generalizations about our world, it is how we make sense of the world. It is how we learn about the world. It is how we talk about the world. But, in the final analysis, all generalizations are at least partially wrong. Each person has their real identity and as we learn about them, we find new things about them. To remember we assign that person additional attributes so that we can remember them and so that we can talk about them. However we model attributes in the digital world, it needs to conform to the use that humans make of attributes in the real world.
Solutions
- Attributes should not be released until informed User Consent is obtained.
References
- NIST Internal Report (NISTIR) 8112: Attribute Metadata defines a schema for metadata that describe a subject’s attributes; it is intended to give relying parties (RPs) greater insight into the methods attributes are determined to assist in making risk-based business decisions. As a result, RPs can examine this metadata and determine if they have the confidence they need in the attribute value before making an authorization decision. This NISTIR is being treated like an “implementers’ draft” – an approach used that focuses on real-world implementation results and lessons-learned before the document can become finalized.
