In an effort to address growing concerns of discrimination and bias in artificial intelligence (AI) algorithms, the Indian government is preparing to introduce provisions in the upcoming Digital India Bill. These provisions aim to enhance algorithmic accountability and grant citizens the right to opt out of decisions made by algorithms. However, opting out may come with conditions, potentially requiring citizens to forfeit certain digital services that rely on algorithmic decisions.
The move comes as AI-based algorithms face scrutiny for their potential biases and discriminatory outcomes. The government’s intention to introduce algorithmic accountability reflects the widespread use of artificial intelligence in various aspects of daily life, including governance.
Under the proposed legislation, the Digital India Bill will replace the two-decades-old Information Technology Act and establish comprehensive oversight mechanisms for the digital landscape. The opt-out mechanism, as outlined in the bill, will allow citizens to decline algorithmic decisions. However, there may be exceptions when the decision is legally required or authorized or if permitted under the law.
The bill is also expected to mandate that organizations provide explanations for algorithmic decisions, including the user characteristics and processes involved. In cases where citizens cannot opt-out, they may have the option to request a human review and reconsideration of the decision. This could have a significant impact on content moderation by social media platforms, which heavily rely on AI automation.
The Digital India Bill is currently undergoing pre-consultations in various cities, with discussions centered on principles such as open internet, online safety, trust, accountability, adjudicatory mechanisms, and new technologies. The government is considering changes, including the potential removal of the safe harbor provision for internet intermediaries, to address the evolving digital landscape effectively.