IMPERATIVE FOR CLEAR ACCOUNTABILITY IN AI SYSTEMS

In 2025, AI advancements may enable attackers to bypass biometric security, posing significant threats to data protection according to the Data Security Council of India (DSCI), set up by Nasscom.

- Ruchira Talapatra

Artificial intelligence (AI) is driving technological revolutions across sectors like healthcare, transportation, law enforcement, and finance. However, as AI systems become integral to these domains, the issue of accountability arises, especially when autonomous systems malfunction or contribute to harmful outcomes. India, as a rapidly growing technology hub, faces the challenge of creating a robust legal and ethical framework to regulate AI effectively.

AI technologies—from machine learning algorithms to deep learning systems—rely on complex data sets to make autonomous decisions. As these systems integrate into daily life, it becomes crucial to assign responsibility for their actions. Accountability is key to ensuring AI's benefits are maximized while mitigating risks such as privacy violations, discrimination, and criminal activity facilitated by AI.

THE "BLACK BOX" PROBLEM

A major challenge in AI accountability is the "black box" nature of many systems, particularly those based on neural networks. These systems can produce outputs that are difficult to trace back to their reasoning, complicating liability assignment when harm occurs. This opacity exacerbates the challenge of holding developers, operators, and users accountable.

THE GAP IN CRIMINAL LIABILITY

Traditional criminal liability concepts, like mens rea (guilty mind), do not apply to AI, which lacks intent or consciousness. This gap calls for a rethinking of how accountability frameworks can evolve to address AI's unique characteristics.

AI AS A LEGAL ENTITY?

Some propose granting legal personhood to AI systems, similar to corporations. This would allow AI systems to hold rights and liabilities, but it risks undermining human accountability. Developers, operators, and users still play crucial roles in programming and deploying AI, and the focus should remain on their responsibilities.