Introduction
Artificial intelligence (AI) is advancing rapidly worldwide, bringing tremendous opportunities and potential risks. Governments scramble to keep up with appropriate regulations as developers race to deploy AI systems. India is taking a significant step forward by proposing the Digital India Act (DIA), a comprehensive regulatory framework governing AI and related technologies. This action mirrors a powerful surge of influence advocating for increased regulation, involving regulatory entities, lobbying efforts, and major tech players.
Here Is What Has Happened So Far
- India lacks a comprehensive legal framework for regulating AI. MeitY has commenced committees that have already released frameworks to standardise responsible and accountable AI development.
- Recently, TRAI, the Telecom Regulator, recommends immediate establishment of regulatory framework that should be applicable across sectors, a move that has gained significance amidst the global AI debate as nations rush to create rules for deploying AI solutions.
- The Data Protection Bill was recently passed and is expected to further catalyze the implementation of AI regulation.
- Flipkart has signaled alarming message to parliament regarding emerging AI threats and supports the newly imposed Data Protection Bill.
Regulating AI: Global Momentum Builds as India Prepares New Rules
The global landscape is witnessing rapid developments as India gears up to take a more definitive stance on AI regulation. The European Union (EU) has taken a proactive step by drafting rules that demand disclosure of AI-generated content, differentiation between deep fakes and genuine content, and prevention of illegal material dissemination. The EU is also pushing for global collaboration on AI, using forums like the G20 to emphasize the need for responsible AI deployment.
In sync with these global efforts, India is introducing the Digital India Tech bill to replace the existing Information Technology Act of 2000. This move is aimed at streamlining and modernizing tech legislation, creating a robust foundation for the future of AI.
Digital India Act: Regulation of AI in E-Commerce and Beyond, not only Generative Models
During the March 9, 2023 presentation by MeitY regarding the Proposed Digital India Act, special attention was given to Bing and ChatGPT. This emphasis aimed to underscore the crucial necessity of establishing protective measures concerning "User Harm," a central aspect of the proposal.
At the heart of the Digital India Act lies the fundamental principle of accountability. The outlined policy places significant importance on two key aspects: algorithmic transparency and the consistent evaluation of risks by digital entities. Central to the legislation are commitments to uphold the constitutional rights of citizens and to foster the responsible and ethical utilization of AI technology.
The Essence of AI Governance and Responsible AI: Model Documentation is Imperative
The evolving regulatory landscape in India indicates that Indian companies need to prioritize developing AI responsibly. Frameworks proposed in the Digital India Act are still taking shape and the government aims to avoid over-regulation that could stifle innovation. However, major stakeholders are pushing for more regulation.
Governmental Bodies and Big Tech Calls for More Regulation
The push for more AI governance and the participation of regulators like TRAI have made prompt AI regulations seem necessary and calls for a more rigorous regulation efforts. Notably, the European Union has deployed representatives lobbying nations like India to emulate their AI approach. Additionally, Flipkart recently warned the parliament about possible increases in AI-enabled cybercrime.
Overall, investing in responsible AI is crucial for Indian firms amidst calls for greater regulation from governmental bodies and industry leaders.
Two Approaches to Meeting Documentation Needs:
As India shapes AI regulation through the Digital India Act, a pivotal challenge emerges: ensuring responsibility and accountability without hampering innovation. With increased influence from industry giants and regulatory bodies, model documentation is the first step toward responsible and governed AI. Two distinct approaches have surfaced, each with its implications.
- Manual Model Documentation
Historically banks have been proponent of this practice. This demands rigorous record-keeping throughout AI model development and has led to a paradigm shift in how models are documented.
Impact:
Manually documenting models comes with a heavy workload and expenses, exemplified in sectors like banking where costs can soar to a substantial
USD 50 million, highlighting its significant impact. - Auto-documentation: Efficiency Amplified
Specialized documentation software like Vectice can automate repetitive documentation tasks. Using Vectice to generate and update model documentation reduces the manual effort required to keep the documentation current.
Impact:
This minimizes the documentation burden and increases the robustness with lower risk and cost.
Regulation Aside: Documentation of AI Efforts is the Right Thing to do
Indian enterprises can secure a competitive edge by adopting responsible AI tools supporting India's specific needs and context.
Comprehensive model documentation stands as a strategic and fundamental aspect for ensuring the ongoing delivery of responsible AI solutions, and partnerships with companies like Vectice, which specialize in auto-documentation for ML projects and their governance, would support this important initiative.
Last updated: August 16th, 2023