Artificial intelligence: Twitter is to work on a more ethical and responsible technology
On April 14, the Californian company presented the first results of its work, with the aim of implementing a more ethical algorithm on the platform, and giving more control to its users.
On its blog, Twitter explains that it wants to improve the machine learning process, i.e. the computer processes that allow machines to automatically learn from the data they are fed.
To do so, the platform relies on a comprehensive team of engineers, researchers and megadata experts. Within this interdisciplinary team is a division dedicated to ethical, transparency and accountability issues in order to understand the undesirable – and sometimes unintended – effects of machine learning.
The social network explains that the use of machine learning can “have an impact on hundreds of millions of tweets per day and that sometimes an algorithm can behave differently from the way it was designed.”
This includes the cropping of photos before publication according to possible sexist or racist bias of the algorithm, whose discovery had caused confusion earlier this year, or the favoring of certain political currents in the distribution of content.
Finally, Twitter wants to give power back to its users so that they can shape their own Twitter. To do this, the company is committed to publishing its research to make the topic understandable and receive comments and feedback from the public.
The social network is therefore committed to implementing a more responsible and transparent machine learning, which, according to the firm, necessarily involves taking responsibility for its algorithmic decisions.
By the IP/IT team of UGGC Avocats
