The ever increasing digitisation has led to the domination of AI in almost every sphere of life, from recruitment to making judicial decisions. Successful implementation of many AI systems remains still remains a challenge due to various issues and algorithm bias is one of them. Organisations need to fundamentally rethink their AI operating models and government interference is of utmost importance. Organisations are more likely to think of their own commercial benefit, therefore, the government interference can ensure the three basic principles advocated by the authors: (i) Transparency (ii) Auditability (iii) Accountability The European Union has already had a governance framework in place which regulates AI and has subsequently released much guidance for a fair deployment of AI. The United States also has an Algorithmic Accountability Act of 2019 in a discussion which is first ever attempting to consolidate principles of algorithmic decision making and regulate AI systems across industries.
The first step towards ensuring a fair deployment of AI is to have a data protection regime in place. In India, data protection is currently governed by the SPDI Rules (as defined hereunder), however, a Personal Data Protection Bill, 2019 (PDP Bill) is also tabled in the parliament and is likely to be passed soon. The PDP Bill does not provide any restriction for decisions solely based on automated processes rather provides the data subject (data principal in the context of the PDP, 2019) the right to receive the personal data which has been processed in a structured and machine-readable manner. BlockSuits recommends an (i) ex-ante approach; and (ii) ex-post approach, to ensure transparency, auditability, and accountability.
Authored by Samaksh Khanna, co-founder and Shivani Agarwal, founder.
Comentarii