PIB reveals report that Center has asked ministries to shun OpenAI, anthropic models of cyber security

The Press Information Bureau (PIB) has dismissed as “false” a media report which claimed that the Ministry of Electronics and Information Technology (MeitY) has ordered government departments to delay the implementation of cyber security models developed by OpenAI and Anthropic.

In a post on X PIB Fact Check, she said the report was incorrect and clarified that no such notice had been issued.

“A report published by @ThePrintIndia claims that the Ministry of Electronics and Information Technology (MeitY) has directed ministries to delay implementation of OpenAI and anthropic models of cyber security,” the post said.

PIB Fact Check rejected the claim, saying, “This claim is false.”

She further clarified: “MeitY has not issued any such direction or recommendation prohibiting ministries from using OpenAI or Anthropic.”

The agency urged people to rely only on official government websites and verified sources for authentic information and appealed to users to report suspicious content to PIB Fact Check to help curb misinformation.

What did the message say?

A report by ThePrint said that the department under MeitY has asked government departments not to implement the cybersecurity models developed by OpenAI and Anthropic for the time being.

According to the report, an office memo was circulated advising ministries not to introduce AI models prematurely. The development reportedly came days after representatives of OpenAI and Anthropic met with officials in several ministries to promote the use of their AI tools. However, the report said it could not independently confirm how many ministries had been approached or the level of meetings.

ThePrint further reported that the Treasury Department has written to the Department for permission to explore the use of agent AI and OpenAI models, including GPT-5.5, for cybersecurity purposes. The department’s six-page letter, titled “In Light of LLMs in Use: AI-Based Vulnerability Discovery, AI-Enabled Cybersecurity Capabilities and Implementations,” outlined possible AI deployments for consideration.

According to the report, the ministry responded with a “firm no” in a memo issued last week. However, a source told the publication that the memorandum should not be considered a permanent ban. Instead, it was about deploying the technology too soon, rather than outright banning its use.

The report states that OpenAI and Anthropic are marketing their advanced AI models for defensive cybersecurity tasks such as identifying software vulnerabilities, inspecting code, resolving security incidents, and performing multi-step tasks with limited human supervision. He also pointed out that these capabilities are considered dual-use, meaning the same technology could potentially be used for both defensive and offensive cyber operations.