Weed out ‘deep fakes’, MeitY tells social media platforms

These platforms have to do so within 24 hours of getting a complaint from an individual, the Ministry of Electronics and Information Technology said in an advisory on Tuesday.

ET has reviewed a copy of the letter, which was also sent to platforms like LinkedIn, Snapchat, and ShareChat.

As per Rule 3(2)(b) of the IT Rules, this content could be impersonation in electronic form, including artificially morphed images of an individual, the letter said.

“This is an advisory. We have been warned of deep fakes by our agencies including some in the MHA, so we have asked the companies to look into it,” MeitY sources told ET.

They added that the ministry expects a prompt response from the companies and will also invite them to discuss different ways in which deep fakes can be restricted.

Discover the stories of your interest

In the email, addressed to the chief compliance officers of these platforms, MeitY highlighted that there were reports regarding the potential use of artificial intelligence-generated deep fakes that were manipulating people by generating doctored content.“…significant social media intermediaries are advised to ensure that their rules and regulations and the user agreement contain appropriate provisions for the users not to host, display, upload, modify, publish, transmit, store, update or share any information that impersonates another person and that the users are duly informed of the same,” according to the email.

Further, intermediaries have been advised to put in place appropriate technology and processes for identifying information that may violate the provisions of rules and regulations or user agreement. It also said expeditious action should be ensured, ‘well within the timeframes stipulated under the IT Rules.’

As per the rules, an intermediary is expected to remove unlawful information or disable access within 36 hours of receiving a court order or being notified by the government or its authorised agencies and within 24 hours when it comes from an individual or a person authorised by the said individual.

In addition, MeitY said intermediaries shall periodically inform its users, at least once every year, that in case of non-compliance with its rules and regulations or user agreement, it has the right to terminate access of usage rights immediately or remove non-compliant information.

Homegrown microblogging platform Koo said impersonation of eminent personalities was one of the five areas which have a high impact on user safety.

“We have developed an in-house ‘MisRep Algorithm’ which constantly scans the platform for profiles using the content or photos or videos or descriptions of well-known personalities. On detection of impersonating content, the pictures and videos are immediately removed, the accounts are immediately flagged for future bad behaviour, and posts are deprecated from circulation. All these actions occur within 10 seconds,” Rajneesh Jaswal, head, legal and policy, Koo told ET.

(Aashish Aryan & Suraksha P contributed to this story.)