CCI’s proposed 'self-audit' framework for AI 'unique' as compared to other jurisdictions: Experts

According to experts, the six-pillar self-audit framework is a significant proposal towards embedding competition compliance into AI governance.

Experts have hailed the Competition Commission of India’s (CCI) new market study’s recommendation of a "self-audit" framework for businesses to ensure that their AI tools comply with the competition laws. They say the proposed "six-pillar" framework is a bold and unique step to ensure compliance by companies, as no other country, including advance economies, has a similar structure in place.

"The six-pillar self-audit framework is a strong step towards embedding competition compliance. While it cannot guarantee the absence of anti-competitive practices, it can substantially reduce risks by institutionalising periodic checks and accountability," said Anisha Chand, Partner at Khaitan & Co.

"⁠At present, UK and US laws encourage internal audits and algorithmic accountability. But, the CCI’s six-pillar model is relatively unique in offering a structured approach and if implemented well, it could become a regional benchmark," noted Chand.

According to the market study, a self-audit mechanism facilitates early identification and mitigation of AI-related competition risks before they cause market harm. It enhances the "auditability and explainability" of AI models. "Therefore, the introduction of a self-audit framework for enterprises deploying AI systems, particularly those with market power or wide consumer reach, will help assess and align their algorithms with competition norms," the study says.

The study defines the six broad principles as follows: (a) Documentation of AI-based decision-making process including algorithmic objectives, data sources, and access protocols; (b) Design and testing of algorithms with built-in safeguards to prevent unintended anti-competitive outcomes and reduce unfair competition risk; (c) Regular internal algorithmic audits/periodic review of algorithmic outputs/market outcomes to proactively identify and eliminate inadvertent algorithm-driven collusion; (d) Review of AI driven pricing strategies to detect unintended price alignments or discriminatory practices; (e) Implementation of safeguards while using third-party tools; and  (f) Incorporation of safeguards that prevent sharing of commercially sensitive data amongst competitors.

According to experts, the six-pillar self-audit framework is a significant proposal towards embedding competition compliance into AI governance. However, its success in curbing anti-competitive practices is likely going to depend on effective implementation and regular internal review by businesses, they say.

"While not a substitute for regulatory oversight, it complements it by fostering a culture of internal accountability and responsible AI deployment," said Parthsarathi Jha, Partner, Economic Laws Practice.

"Much like the self-regulatory framework proposed in the CCI’s 2020 market study on e-commerce, this self-audit mechanism serves as a proactive compliance tool, complementing rather than diluting the CCI’s enforcement powers to ensure prevention of anti-competitive practices," Jha added.

Shivanghi Sukumar, Partner at Axiom 5 law chambers said that few jurisdictions have formal self-audit frameworks specific to AI and competition law, which makes the CCI’s approach relatively distinctive. "The CCI’s model sets out a process-oriented framework that businesses can use to evaluate and document their own AI systems. In other words, it moves from what to consider to how to act," she said.