Knowledge creation by AI

In recent documentation produced by the TV program ARTE the tory of Wikipedia was presented.

As a frequent user of Wikipedia I generally trusted the information provided on the website. Now my awareness for continued checking of the provided information significantly changed. As I have learned in the documentation, an automated article writing program (the LSJBOT) was developed to generate new entries in Wikipedia. The focus of the automation was on nature and animals and its presentation in the languages Swedish, Caray and Cebuano. With a capacity to create up to 10.000 new article per day, the entries in the mentioned languages quickly surpassed a majority of entries in other languages.

From my perspective there are several perspectives to be considered. First, it is a positive approach, when subject-related information become available in "exotic" mother languages like Caray or Cebuano. But this does not directly correspond to the commonly used and required information in a daily life context.

Secondly, in case of the bot mechanism, strict rules had been applied to produce standardized entries, just differing in the subject. Here an automation process upscaled a human writing process based on big data analysis and result presentation.

Thirdly, the approach raised my awareness in general to review and check Open Educational Resources (OER) much more for their correctness and if other concepts are existing.

Fourthly, I get increasingly the impression, that digital competences should include the understanding how AI works, especially concerning the different methods of machine learning.