As artificial intelligence becomes embedded across enterprise processes and public systems, the defining challenge will not simply be how quickly organisations adopt the technology, but whether users can continue to trust the systems and decisions it produces, according to Dr. Amar Patnaik, former member of Parliament, Rajya Sabha.
Delivering his Expert Remarks at the ETLegalWorld’s AI-Powered Legal Transformation Summit 2026, Patnaik said trust will remain the constant linking technological innovation, regulation, enterprise governance and consumer confidence.
Drawing on his experience across civil services, policymaking, parliamentary committees and legal practice, he said the next phase of AI governance must be viewed from two interconnected perspectives: how lawmakers respond to technological change and how enterprises redesign their processes to comply with emerging laws.
“Whatever we do, whatever regulations that we bring in, whatever interpretation of law… the variable which is going to remain cemented permanently is trust,” Patnaik said.
Why India is leaning towards principle-based regulation
One of the biggest regulatory challenges, Patnaik said, is that technological development typically moves faster than legislation. That reality, he argued, strengthens the case for principle-based and technology-agnostic regulation rather than laws that attempt to prescribe rules for every technological development.He pointed to India’s shift in data protection legislation towards a more principle-based framework, arguing that fundamental principles should sit within legislation while more detailed prescriptions can be addressed through rules, guidelines and advisories that are comparatively easier to update.
Patnaik said India has so far avoided adopting an AI law comparable to the European Union’s framework, instead relying more on guidelines and advisories while attempting to balance digital growth with safeguards.
For enterprises, Patnaik said responsible AI must begin with lawfulness and privacy by design. Organisations should evaluate whether the data they collect is genuinely necessary for delivering a particular product or service and whether the purpose for which it is being processed remains aligned with the consent originally obtained.
Patnaik highlighted principles including purpose limitation, collection limitation, auditability, accountability and data quality as elements that enterprises will need to build into their systems and processes.
AI creates an explainability and auditability problem
Among the most significant governance challenges created by AI, Patnaik identified the potential loss of both audit trails and explainability.
That poses a problem when an organisation later needs to demonstrate why data was processed in a certain manner, whether that processing remained within the original purpose for which consent was obtained, or who was responsible for a consequential automated decision.
Patnaik said legal leaders will increasingly need to examine whether data-minimisation principles are being followed, whether AI systems preserve adequate logs, and whether enterprise processes remain capable of being explained when scrutinised by auditors, regulators or courts.
“Explainability becomes an issue, auditability becomes an issue,” he said, warning that when audit trails disappear, enforcing accountability becomes considerably more difficult.
Someone must remain accountable for an AI decision
That challenge leads to what Patnaik described as one of the most fundamental requirements for credible AI governance: a human must ultimately remain answerable.
The issue becomes more complex as organisations experiment with agentic AI systems capable of taking or recommending actions with decreasing levels of direct human intervention. An AI agent itself cannot easily become the accountable legal entity when harm occurs.
For every consequential automated decision, Patnaik argued, organisations should be able to identify a lawful basis for the action, explain the reasoning behind it, identify a named human owner, maintain an audit record and provide a mechanism through which affected parties can seek redress.
Patnaik described the general counsel as potentially becoming a form of “corporation’s constitutional officer”, responsible for helping ensure that automated enterprise decisions remain within the boundaries of law and accountability.
Boards and CXOs will need new AI governance capabilities
Patnaik also emphasised that rules alone will not create responsible AI ecosystems.
Institutions will require the capacity to evaluate AI models, enforce regulatory requirements, establish accountability and respond effectively when individuals or companies suffer harm. That capability must extend to the highest levels of corporate leadership.
Audit committees, boards and C-suite executives, he said, will need greater understanding of AI governance because responsibility for technology-related failures will increasingly reach senior decision-makers.
“Training at the highest level” will therefore become critical to building institutional capacity around AI, Patnaik said.
As AI grows more autonomous and becomes embedded deeper into corporate decision-making, the organisations best positioned to sustain adoption will be those that can demonstrate not only what their technology can do, but also who remains accountable, how decisions can be explained and what happens when the system gets it wrong.
Today, the AI-Powered Legal Transformation Summit 2026 by ETLegalWorld has brought together general counsel, in-house legal leaders, and technology innovators, in Bengaluru, to chart the next phase of legal transformation in an AI-driven world.
