IndiaAI Mission Promotes 8 Projects to Foster Ethical AI Development

IndiaAI Mission Favour 8 projects to nurture ethical AI Establishment

The IndiaAI Mission has recently announced its selection of eight projects that seek to promote ethical development of artificial intelligence (AI) in India. This initiative is coupled with a larger government commitment-òn of INR 10,738 crores- to the promotion of responsible AI practices in India.

Key highlights of the selected projects

Machine Unlearning

Lead Institution: Indian Institute of Technology (IIT) Jodhpur

Purpose: Development of techniques that allow AI models to "forget" specific data points, enhancing privacy and fairness.

Synthetic Data Generation

Lead Institution: IIT Roorkee

Purpose: The generation of synthetic datasets to mitigate the bias in machine learning processes.

AI Bias Mitigation in Healthcare

Lead Institution: National Institute of Technology (NIT) Raipur

Purpose: Strategies to reduce biases in AI application within healthcare systems.

Explainable AI Framework

In collaboration with: Defence Institute of Advanced Technology (DIAT), Mindgraph Technology

Aim: To create explainable and privacy-preserving AI models for security applications.

Privacy-Enhancing Machine Learning Models

In collaboration with: IIT Delhi, IIT Dharwad, IIIT Delhi, Telecommunication Engineering Center (TEC)

Aim: Creation of user-privacy-preserving conditions for models that maintain a certain level of performance.

Fairness-Assessment Tools for AI Models

Lead Institution: Indraprastha Institute of Information Technology (IIIT) Delhi

Purpose: Tools to assess the fairness and ethical standards of AI models.

Algorithm Auditing Toolkit (ParakhAI)

Lead Institution: Civic Data Labs

Purpose: An open-source toolkit developed to audit AI algorithms and provide platforms designated for accountability.

Transparency and Risk Assessment Framework for Large Language Models (Track-LLM)

Collaborators: Amrita Vishwa Vidyapeetham, TEC

Aim: The framework whose objective is to renew contacts with large language model-related risk areas.

Context & Important Comments

The projects are selected among 2,000 plus proposals, which reflects the Indian government's strong commitment to issue remedial measures for issues such as bias mitigation, privacy concerns, and establishing governance frameworks for AI technologies. The aim of the initiative is aligned with the government's plan of creating indigenous tools and frameworks to address India's unique challenges in the AI space.