In a move to tackle the growing threat of deepfake technology, Union Minister for Electronics and Information Technology and Communications, Ashwini Vaishnaw, chaired a high-level meeting on Thursday. The meeting was attended by the representatives of major social media platforms, companies making AI tools, NASSCOM and professors from the field of AI.
Describing deepfakes as a emerging threat to democracy, the IT Minister announced that the government is set to introduce draft regulations, which will be open to public consultation, within next 10 days to address the issue. The minister stated that companies have acknowledged the necessity for decisive action in key areas, including detection, prevention, fortifying reporting mechanisms, and enhancing user awareness.
Key highlights from the meeting
- Addressing the media, post the meeting, Vaishnaw elaborated on the government's plans of taking actions based on 'four pillars'. These four pillar include detection of deepfake, prevention of their spread, strengthening the reporting mechanism and awareness. He added that today we have tools which work within the purview of privacy and can effectively detect deepfakes.
- With the regulations coming in, the accountability will lie not only on the person who creates the deepfake and spreads it but also on the platform that hosts that deepfake. Vaishaw stated the these regulations can come either in the form of a new law or amendments to the existing laws.
- According to the IT minister, labelling and watermarking deepfake is an absolute necessity but not enough. "All the social media platforms agreed on it, but we need to do much more," he said.
- Vaishnaw told media that all the social media platforms came to the consensus that deepfake is a very big threat to the society. "They all are already taking steps internally and are looking to increase the intensity of those efforts," he said. The platforms have also been asked to come up with their thoughts on how to to tackle the apps which facilitate in creating deepfake videos.
- Talking about deepfake videos which may not originate from India but is circulated within the country, Vaishnaw said that anything which is created anywhere but is being circulated within India's geography will be regulated.
- Addressing a question about distinguishing between deepfakes and misinformation, Vaishnaw said that all of the synthetic content cannot be termed harmful. "Somebody using AI simply to enhance the photograph is not harmful. That's not something which is affecting the society but there is somebody who is using somebody's face and creating a video can be totally wrong. For example, in the recent elections in Madhya Pradesh a video surface in which the Chief Minister was kind of saying that you can vote for the opposition party. That is a deep misinformation," he said.
Deepfakes' Impact on the Political Sphere
According to Vaishnaw, deepfakes and AI-generated synthetic content are undermining free speech and privacy. Responding to a media inquiry about the Election Commission's involvement in tackling the issue, and whether the IT ministry plans to engage with political parties to obtain a fundamental commitment to abstain from using deepfake technology to portray opposition leaders misleadingly, he said, "We are open to all consultation from anybody who would like to point out their concern in the public consultation round."
On the question of whether the government would allow, a political party or a political party campaigner to make a video in different languages to reach out to different demographies, the IT minister, without giving an affirmative answer said that the regulations will aim at giving users right to know "what is natural and what is not natural".
As per the meeting, the IT ministry aims to come with regulations within 10 days from now, but deepfakes have already stepped into the political communications, ahead of the election season. This technology is already being used by Divyendra Singh Jadoun's company Polymath Solutions to create "personalised messages" for Congress party workers involved in the preparations for the forthcoming Rajasthan assembly elections. The company will also undertake the similar task for BJP, before the Lok Sabha elections of 2024, he told Decode.
Not just political messaging, the menace has already made its place in spreading misinformation. Back in 2020, in the first ever use of AI-generated deepfakes in a political campaign, a series of videos of Bharatiya Janata Party leader Manoj Tiwari appeared on multiple WhatsApp groups. The videos showed Tiwari hurling various allegations to his rival Arvind Kejriwal in English and Haryanvi, before the Delhi elections.
In a similar such example of a recent viral video, former chief minister of Madhya Pradesh Kamal Nath is seen addressing a crowd and speaking about cancelling the Laadli Behna Scheme. Another instance include several doctored videos with fabricated voiceovers, of the game show Kaun Banega Crorepati, shared by pro-Congress social media accounts. These are all examples of depfake videos.