Twitter has announced that it will award rewards to users and researchers who discover potential biases in its platform´s algorithms, linked, for example, to sexism or racism.
This competition is the first in the topic, according to Twitter, which indicated that prizes may amount to 3,500 dollars.
This idea is reproduced from competitions offered by some websites to discover gaps in information security, according to what the two officials in the network, Roman Choudhury and Juta Williams, explained in a message.
"It´s hard to find biases in machine learning models, and sometimes companies find they have inadvertently broken ethics after using" these models, the two officials said, "and we want that to change."
In the developed model of vulnerability detection, researchers and hackers alike helped IT security managers “create best practices for identifying and managing vulnerabilities in order to protect the public,” Roman Choudhury and Juta Williams said.
"We want to develop a similar community" to detect algorithmic biases, the two Twitter officials added.
In April, Twitter presented its ongoing work to make the algorithms behind the scenes of the platform more ethical, a way to respond to criticism about the risks associated with these technologies.
The social network ditched a photo algorithm a few weeks later after finding it was slightly biased in favor of white people.
Twitter is offering rewards to anyone who discovers biases in its algorithms
