A draught standard for evaluating and measuring the fairness of artificial intelligence systems was published by the convergence and broadcasting division. Applications of machine learning (ML) and artificial intelligence (AI) are being employed more and more in all fields. Unintentional biases in their forecasts or results are a major worry.
In order to ensure that the fairness assessment is customised to the appropriate use case, the cooperation investigated the dynamics linked to fairness in the Indian environment, including the diversity, the protected and proxy factors, and approaching fairness in a risk-based manner. Additionally, it makes an effort to account for numerous sources of bias using a scenario-based approach that is process, metrics, and adversarial/causality-based.
The AI/ML System’s impartiality or fairness is a key component of responsible AI. In order to fulfil the Government of India’s goal of increasing public confidence in AI/ML systems (#AIforAll), TEC has started a voluntary fairness assessment of these systems.
A working group comprised of representatives from industry, academia, research, subject matter experts from government agencies, and others was formed by TEC as a result of the public consultations to create the first draught of the proposed Standard for evaluating the fairness of AI Systems.
In order to frame procedures for evaluating fairness for various types of AI/ML Systems and for issuing fairness ratings/certifications to AI/ML systems, as a benchmark of fairness, C&B Division TEC held public consultation through an interactive webinar on March 22, 2022, followed by a consultation meeting on September 1, 2022.
