In the presentation of its ongoing work, Twitter has highlighted, this April 14, the ethical AI. The social network wants to improve its machine learning algorithms in such a way that they are more ethical and transparent for people registered on the platform. The goal is to give a little more control to users while responding to criticism related to the potential dangers that these technologies can cause.

A transparent and ethical approach

Social networks, including of course Twitter, are closely watched by NGOs and politicians sensitive to the issue of data protection and the ethics of AI. The founder and CEO of the social network, Jack Dorsey, has repeatedly expressed his desire to give more power to users and to make content moderation more “decentralized.

The company said of the move:

“Algorithmic choice will give users more control to shape the Twitter they want. We are in an exploratory stage on this.”

Studies to come

Twitter says it is working on changing its machine learning algorithms to ensure “they take responsibility and transparency for their decisions on the front end and fairness of results on the back end.” This follows Jack Dorsey’s comments in February suggesting that users could rank algorithms based on what they want to see.

The research team that will be working on changing these algorithms will have to publish several studies regarding possible racist or sexist aspects of the algorithm that, for example, trim images before they are published. This study will also look at the hot topic of content recommendations according to political ideologies in seven countries.

Translated from Twitter annonce des travaux visant à rendre ses algorithmes plus éthiques et transparents