IT platform to the rescue!Government lifts licensing requirements for untested AI models

The government has lifted licensing requirements for untested AI models but stressed the need for labeling of AI-generated content, according to the latest consultation on artificial intelligence (AI) technology.

The new advisory issued by the Ministry of Electronics and Information Technology on Friday evening did not sanction the artificial intelligence model under development but fine-tuned the compliance requirements under the IT Rules 2021.

“This advisory was issued to suppress… a notification dated March 1, 2024,” the advisory said.

According to a new advisory report, it has been found that IT companies and platforms often neglect to perform due diligence obligations highlighted in the Information Technology (Intermediary Guidelines and Digital Media Code of Ethics) Rules, 2021.

The government has required companies to label content generated using their AI software or platforms and to inform users that the output generated using their AI tools may contain inherent errors or unreliability.

“Any intermediary that allows or facilitates the synthetic creation, generation or modification of text, audio, video or audiovisual information through its software or any other computer resources in such a way that such information may be used as misinformation or deepfakes is advised to or any other computer resource created or modified such information is marked as…such information was created, generated or modified using the intermediary’s computer resources,” the advisory states.

It added that if a user makes any changes, the metadata should be configured to identify the user or computer resource that affected such changes.

After controversy erupted over Google’s AI platform’s response to Prime Minister Narendra Modi’s query, the government on March 1 issued an advisory to social media and other platforms to label and label AI models being trialed. Prevent hosting of illegal content.

In an advisory to intermediaries and platforms, the Ministry of Electronics and Information Technology warned of criminal action if regulations are not followed.

Previous consultations required entities to seek government approval to deploy artificial intelligence (AI) models that were experimental or unreliable and only do so after they were flagged as “the output generated may be and inherently erroneous or unreliable”.