Pablo M. Flores
banner
pablomflores.phd
Pablo M. Flores
@pablomflores.phd
PhD @ucdavis.bsky.social | Computational Social Science, Complexity, Networks, Emotions |
Para mí, amor en 48 compases 🎶 🇨🇱
Actualizado!
November 15, 2024 at 10:09 PM
We also detected the existence of spirals of "bad" recommendations (down the rabbit hole). Yet, the likelihood of spirals of "good" recommendations is even larger than the bad ones.
If interested, you can take a look at the full paper available for open access.
#commsky #csssky #algorithmaudits
January 16, 2024 at 10:28 PM
After analyzing 151 algorithm audits from 33 studies, we found that roughly 8-10% of algorithmic recommendations are "bad" while about a quarter actively protect users from self-induced harms ("do good"). This average is consistent across audits irrespective of the platform or the kind of risk.
January 16, 2024 at 10:28 PM