Use AI to filter social media, says government

Asks platforms to meet requests to take down objectionable posts in four hours

Flat line design website banner of web search, SEO, ranking sites, rating. Modern vector illustration for web design, marketing and print material.

The Union government has asked social media platforms to emulate the practices they follow in the U.K. in using tools based on Artificial Intelligence to identify and take down content that incites people to commit violence, a senior Home Ministry official said on Friday.

On a request was made by a law enforcement agency, the Ministry has reduced the time limit to take down such posts and videos to about four hours from about 36 hours at present.

The Ministry has asked Twitter to engage “flaggers” and “voluntary organisations” to identify, detect and take down objectionable content in real time.

“While we recognise the utility of social media platforms, government cannot remain a mute spectator and has to ensure that they comply with the directives. They cannot take ages to respond to requests by agencies to bring down such content,” said the official.

The official said that after the government held several meetings with representatives from Google, WhatsApp, Twitter, Facebook and others, the compliance rates in honouring requests by law enforcement agencies have gone up from 60% to 80%.

“Google has agreed to deploy pro-active tools to voluntarily identify and remove such content. Social media platforms have been asked to engage NGOs working in the field, and appoint a grievance redressal cell, for prompt action,” said the official.

The Ministry had asked Twitter to produce telephone numbers and other identifiable sources that could help track suspicious accounts, whenever asked for by law enforcement agencies.

The Centre has stepped up its interactions with social media platforms to arrest the spread of rumours, and messages inciting unrest, cybercrimes and other activities that could be detrimental to national security. In May and June, over 20 people have been lynched based on fake posts or rumours floating on WhatsApp.

WhatsApp has been asked to keep the complete trail of forwarding of unlawful content and share “meta data” with law enforcement agencies to identify the user who “originated the content.”