The amount to which entities are duplicated affects their acceptability in online communities. Knowledge entities that are heavily duplicated (e.g. the same file as another knowledge entity) are nearly universally discouraged, while two or more similar entities covering the same topic in their own way but with overlapping information may be acceptable in some online communities. 

Spectrum of Duplication and Acceptability in Online Communities

Our reliance on search engines and language models to curate the knowledge we are exposed to can pose a significant threat to society. The power wielded by tech giants to direct individuals towards their desired objectives could have a profound impact on the future of societies.

Search engines and language models can manipulate our future

Searching for learning can be highly inefficient as search results often assume prerequisite knowledge that the reader may not possess. Such results may include articles, blog posts, videos, and other content that assumes a certain level of understanding, making it challenging for the reader to learn effectively.

Inefficient “Searching for learning”

Connecting the knowledge acquired from courses, books, and other sources can be challenging as authors and instructors often focus on making their material coherent within the scope of their work, disregarding its coherence in relation to other courses, books, or resources that the reader may have encountered earlier.

Disconnected knowledge entities in Online Communities

The dissemination of misinformation and disinformation can be facilitated by many factors, such as mass dissemination and burying it within factual and popular information, making it easier to go unnoticed or accepted as true.

Vulnerability to Misinformation and Disinformation in Online Communities

- Duplicate knowledge entities
- Overlapping knowledge entities
- Increased dissemination of misinformation and disinformation
- Search engines and language models can manipulate our future
- Inefficient “Searching for learning”
- Disconnected knowledge entities

University of Michigan - Ann Arbor

The exponential increase in Internet publications, coupled with the emergence of extensive language models like ChatGPT, has resulted in an overwhelming knowledge overload.

Learn Before

Related

Learn After