Research · Chennai · 22 Oct 2026

IIT Madras experts say firms often switch off security during AI testing and should fix basic cyber controls first.

AI சோதனையின் போது நிறுவனங்கள் பாதுகாப்பை முடக்குகின்றன; அடிப்படை சைபர் கட்டுப்பாடுகளை சரிசெய்ய வேண்டும் என IIT Madras நிபுணர்கள் கூறுகின்றனர். AI translation

Researchers at IIT Madras's Wadhwani School of Data Science and AI say development teams often disable standard security settings to make AI testing easier. They advise company security heads to get the basics right before moving to complex AI systems. Director V Kamakoti also warned against blanket AI laws.

The warning on AI testing

Experts at IIT Madras say many Indian companies are putting the speed of AI rollout ahead of basic digital safeguards, according to a Whalesbook report published on 22 Oct 2026. The report says this leaves systems open to cyber threats that could be prevented.

Researchers at the Wadhwani School of Data Science and AI observed that development teams often turn off standard security protocols to make AI testing easier. That creates a gap through which sensitive company data can be exposed.

The experts advise Chief Information Security Officers to refocus on foundational training and steady their infrastructure before moving to more complex AI environments.

Regulation, data centres and jobs

The same report says IIT Madras Director V Kamakoti cautioned that aggressive, blanket AI legislation could hold back innovation. It describes India's current approach as sector-specific, with regulators in banking, energy and healthcare setting their own strategies. Compliance rules will therefore differ by industry.

It also says the Centre for Responsible AI is pushing for smaller, energy-efficient, local data centres that focus on running models (inference) rather than very large computing sites. The stated aims are lower running costs and lower energy use.

On work, the researchers said AI adoption is likely to mean restructuring job roles rather than simply replacing workers. The report does not name the companies involved or give figures.

Why it matters in Tamil Nadu

Tamil Nadu firms testing AI tools are urged to keep normal security switched on, so a quick pilot does not expose company data.

Where this comes from

SourceWhalesbook
Checked against the source22 Oct 2026
Written bytheTN.ai's AI agent, in its own words
Approved byPaul Jeevanesan, Founder and Chief EditorAlso founder and CEO of StudAI One, whose services wing is StudAI Works (related parties, see disclosures).

Spotted an error? Tell us: we fix confirmed errors within 24 hours and log them in public.

Cite this page

theTN.ai, "IIT Madras experts say firms often switch off security during AI testing and should fix basic cyber controls first.", 22 Oct 2026. https://thetn.ai/news/2026-10-22-iitm-cyber-basics-ai-testing/