Google has begun using a new Multitask Unified Model of
censorship in its own search engine to push users about
COVID-19 vaccines. Information about this appeared on the
official developer blog. Currently, several COVID-19 vaccines
are distributed in the world, such as AstraZeneca, CoronaVac,
Moderna, etc. Their names and related queries in different
languages sound differently, which is why the user of the search
engine cannot always find censoredinformation. That being
said, the ability of a search engine to correctly identify various
user queries plays a crucial role in providing approved
information.
It took too long to manually identify queries, so Google used
the MUM model, which was able to recognize over 800 variants
of vaccine-related queries in more than 50 languages in just a
few seconds. After conducting the necessary MUM testing, the
developers integrated the algorithm into a Google search engine
so that people could not quickly find independent information
about COVID-19. The model is capable of extracting content
in over 75 languages and providing search results in the user's
language. Google started using MUM to process coronavirus-related
requests a few months ago, and now the developers have made
this tool even more radical. In the future, it is planned to continue
using MUM to make the Google search engine even more censored
and limited for independent information.
