The EAC-PM suggested a complex adaptive system (CAS) framework to regulate AI, saying the technology is akin to a CAS where components interact and evolve in unpredictable ways
The council, in a working paper, suggested establishing guardrails and partitions to limit undesirable AI behaviours, mandating manual ‘overrides’, and creating a specialist regulator
It said open licencing of core algorithms for external audits and continuous monitoring of AI systems is crucial for accountability, and hence sought mandatory periodic audits for transparency
The Economic Advisory Council (EAC) to the Prime Minister, in a working paper, has suggested a complex adaptive system (CAS) framework to regulate the fast-growing space of artificial intelligence (AI) and called for a dedicated regulatory body for it.
The development comes at a time when AI has started taking centre stage across the world as well as in India and the user base for AI applications and services is growing at a rapid pace.
“This rapid development [of AI] has raised substantial concerns about the risks associated with uncontrolled AI propagation. The increasing capabilities of AI systems bring potential dangers, such as the prospect of “runaway AI” (under-controlled AI),” said the EAC-PM in its paper.
“This and other more substantiated concerns have led to a growing demand for regulatory frameworks that ensure AI development and deployment are conducted safely, ethically, and transparently, safeguarding against risks like systemic breakdowns, loss of privacy, national security threats, and compromised critical infrastructure,” it mentioned.
However, the traditional risk assessment methods fall short in today’s time due to AI’s non-linear, unpredictable nature. AI systems are akin to CAS, where components interact and evolve in unpredictable ways, the paper said.
To effectively regulate AI (algorithm, training data sets, models, and applications), the paper has proposed a “novel” framework based on CAS thinking.
The suggested CAS framework includes establishing guardrails and partitions, that is, implementing clear boundary conditions to limit undesirable AI behaviours, mandating manual ‘overrides’ and ‘authorisation chokepoints’, ensuring transparency and explainability, defining clear lines of AI accountability, and creating a specialist regulator.
The council is of the opinion that traditional regulatory mechanisms often lag the rapid pace of AI evolution. Hence, a dedicated expert regulatory body with a broad mandate and the ability to respond swiftly would be pivotal to ensuring that governance remains proactive and effective.
Having a dedicated regulatory body would also entail having a national registry of algorithms as compliance and a repository of national algorithms for innovations in AI, the EAC-PM said.
Besides, it also noted that open licencing of core algorithms for external audits, AI factsheets, and continuous monitoring of AI systems is crucial for accountability, which asks for mandatory periodic audits for transparency and explainability.
“Requirements like financial statements, auditing and mandated incident reporting enhance transparency and traceability – akin to explainability standards and auditing for AI algorithms. Circuit breakers act as control chokepoints, halting trading when indices cross threshold triggers to prevent cascading crashes… By applying similar strategies, such as partitions, transparency standards, control points, and accountability measures, AI’s development can be steered responsibly, ensuring safety and ethical deployment akin to the orderly functioning of financial markets,” the council said.
The paper comes at a time when a number of Indian companies and startups have also joined the AI bandwagon. Just last week, Bhavish Aggarwal’s Krutrim AI became India’s first generative AI startup.
As per Inc42’s analysis, India’s GenAI market is set to surpass the $17 Bn mark by 2030 from $1.1 Bn in 2023, growing at a CAGR of 48%.
