Archive · 4 Oct 2026
Experts Urge India to Build Its Own AI Safety Framework
Experts say India should make AI safety rules that fit the country instead of copying another country's plan.
They say the rules should account for India's many languages and the way AI is used to build applications.
Some experts want laws that require companies to follow safety rules, while others favor voluntary steps to protect innovation and competition.
Several support combining required rules for high-risk uses with flexible guidance in fast-changing areas.
They recommend testing AI, independent checks, and clear responsibility when serious problems occur.
They also say the rules should involve government, companies, researchers, workers, users, and civil society.
The Indian government is considering whether new laws are needed.
Concerns include fake videos, cyberattacks, election interference, and harm to essential services.
Industry experts say India's AI safety framework should reflect its languages, scale, and application-focused ecosystem rather than copy the US voluntary pact.
The White House Accord on Super Intelligence was signed September 29 by Donald Trump and executives from six technology companies.
Several experts favor binding safeguards or a hybrid model; Gnani.ai CEO Ganesh Gopalan argues mandatory rules could hinder competition and favors self-regulation.
Proposed measures include system testing, independent audits, serious-incident reporting, accountability, and stronger oversight of high-risk uses.
Experts cite risks including deepfakes, cyberattacks, electoral interference, privacy breaches, and harm to critical infrastructure.
- Who
- Indian industry experts, government officials, and the companies that signed the White House Accord on Super Intelligence.
- What
- Experts called for an India-specific AI safety framework and debated voluntary, binding, and hybrid approaches.
- When
- The accord was signed September 29; the PTI report was dated October 4.
- Where
- India; the framework is being discussed in comparison with a US-led voluntary pact.
- Why
- Experts say India's languages, scale, application-led AI ecosystem, and potential AI harms call for a framework suited to the country.
Supporters of binding or hybrid rules
Supporters of voluntary self-regulation
How safety rules should be enforced
Supporters of binding or hybrid rules
Srikanth Velamakanni says voluntary measures alone are insufficient and calls for binding rules on transparency, incident reporting, and accountability. Soumen Mandal and Amit Khanna favor hybrid approaches, including mandatory safeguards in higher-risk areas.
Supporters of voluntary self-regulation
Ganesh Gopalan favors voluntary self-regulation, arguing that mandatory regulation could stifle competition and give some companies preferential treatment.
Balancing safety and innovation
Supporters of binding or hybrid rules
Supporters of binding or hybrid rules argue that oversight and safeguards can address serious risks while allowing flexibility in fast-moving technical areas.
Supporters of voluntary self-regulation
Gopalan argues that India needs more frontier AI companies and that mandatory rules could hamper competition.
- White House accord
- The voluntary White House Accord on Super Intelligence was signed September 29.
- Named signatories
- Google, OpenAI, Anthropic, Meta, Nvidia, and xAI.
- Accord commitments
- Internal model controls, oversight teams, independent external audits, board oversight, and regular safety-standard meetings.
- Approaches debated
- Voluntary self-regulation, binding rules, and hybrid approaches.
- Suggested safeguards
- Testing, independent audits, serious-incident reporting, transparency, and accountability.
- Risks raised
- Deepfakes, fraud, privacy breaches, cyberattacks, electoral interference, and attacks on critical infrastructure.
- Government position
- Officials have said additional AI regulation or a new legal framework may be needed.
