RBI Governor Shaktikanta Das Stresses Need To Eliminate Algorithm Biases

File Photo

New Delhi: RBI Governor Shaktikanta Das on friday stressed the importance of eliminating biases in algorithms because of the increasing use of artificial intelligence (AI) and machine learning (ML).

Governor Shaktikanta Das while speaking at the 18th Statistics Day Conference organised by the Reserve Bank of India (RBI) highlighted the growing use of statistics as a preferred tool for making inferences in various fields. He further noted that the discipline has evolved from simply collecting facts to focusing more on interpreting data and drawing conclusions, especially considering the level of uncertainty involved. ()

"Looking ahead, the year 2025 has a special significance for compilation of official statistics the world over. Global efforts are expected to culminate in new global standards for compilation of macroeconomic statistics, especially for national accounts and balance of payments," he said and added that the team at the Reserve Bank is closely monitoring these developments. ()

He further said RBI is making efforts to harness the availability of huge computing power and growing digital footprints to analyse measures of expectations, sentiment indicators and policy credibility measures from alternative data sources. The governor emphasised that the volume of digital data stored as well as the storage capacity are growing at an exponential pace, bringing forth new challenges along with new opportunities.

"The focus now is naturally on enhancing capacity in artificial intelligence (AI) and machine learning (ML) techniques and analysing unstructured textual data. While doing so, ethical considerations need to be addressed and biases in algorithms need to be eliminated," he said.

The Reserve Bank has ventured into AI/ML analytics in multiple areas. Under the Reserve Bank's aspirational goals for RBI@100, he said the central bank is aiming to develop cutting edge systems for high frequency and real-time data monitoring and analysis. (With PTI Inputs)