FairCVtest Demo: Understanding Bias in Multimodal Learning with a Testbed in Fair Automatic Recruitment

With the aim of studying how current multimodal AI algorithms based on\nheterogeneous sources of information are affected by sensitive elements and\ninner biases in the data, this demonstrator experiments over an automated\nrecruitment testbed based on Curriculum Vitae: FairCVtest. The presence of\ndecision-making algorithms in society is rapidly increasing nowadays, while\nconcerns about their transparency and the possibility of these algorithms\nbecoming new sources of discrimination are arising. This demo shows the\ncapacity of the Artificial Intelligence (AI) behind a recruitment tool to\nextract sensitive information from unstructured data, and exploit it in\ncombination to data biases in undesirable (unfair) ways. Aditionally, the demo\nincludes a new algorithm (SensitiveNets) for discrimination-aware learning\nwhich eliminates sensitive information in our multimodal AI framework.\n