PIB debunks report claiming Centre asked ministries to avoid OpenAI, Anthropic cybersecurity models

The Press Information Bureau (PIB) has dismissed as "fake" a media report claiming that the Ministry of Electronics and Information Technology (MeitY) directed government ministries to hold off on deploying cybersecurity models developed by OpenAI and Anthropic.

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

"A news report published by @ThePrintIndia claims that the Ministry of Electronics and Information Technology (MeitY) has directed ministries to hold off on deploying OpenAI and Anthropic cybersecurity models," the post said.

Rejecting the claim, PIB Fact Check stated: "This claim is fake."

It further clarified, "MeitY has not issued any such direction or advisory 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 report say?

A report by ThePrint stated that a department under MeitY has asked government ministries not to deploy cybersecurity models developed by OpenAI and Anthropic for the time being.

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

ThePrint further reported that the Ministry of Finance had written to the department seeking permission to explore the use of agentic AI and OpenAI's models, including GPT-5.5, for cybersecurity purposes. The ministry's six-page letter, titled “In light of LLMs being used: AI-based vulnerability discovery, AI-assisted cybersecurity capabilities and implementation”outlined possible AI deployments for consideration.

According to the report, the department responded with a "firm no" in a memorandum issued last week. However, a source told the publication that the memorandum should not be viewed as a permanent ban. Instead, the concern was about deploying the technology too early rather than prohibiting its use altogether.

The report noted that OpenAI and Anthropic market their advanced AI models for defensive cybersecurity tasks such as identifying software vulnerabilities, reviewing code, handling security incidents and carrying out multi-step tasks with limited human supervision. It 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.