We have been discussing the problem of “Fake News” in India particularly in the context of the forthcoming Indian elections. The political party in opposition has made it it’s policy to try and win the election only by brazen lies being spoken off without any hesitation under the assumption that some of the mud thrown will stick on their political opponents. To support such world of lies, Internet is being used freely and this needs to be recognized and checked before the entire Internet becomes completely untrustworthy.
Articles being planted in the media by bribing journalists is an old trick. Today, the political parties manage a laboratory to create fake news and spread it across the social media through the millions of fake Twitter or Facebook accounts that are created only for this purpose. One of the tools they use is “Artificial Intelligence” to create news stories that are created to suit their own narrative without any reference to the truth.
Today, even the illiterate rural person knows that TV news is like reality shows. Take it if you like and reject it if you don’t like. People over a course of time have developed an instinct to create his own filter to believe or reject news stories even if the news anchors think that they are successfully brainwashing the public.
Use of “Morphed pictures” was the next tool that fake news creators started using to prove their point. Then they started manipulating the audio stream in a video to change what a video was supposed to show as in the case of the JNU campus incident.
Just as we thought we have reached the end of the technology of fake news creation, comes the alarming news about “Deepfake videos”. (Refer here).
Deepfake videos are created by the advanced use of Artificial Intelligence (AI) where fake videos are created of persons from the machine learning that takes place by observing some real videos. The improvement of the algorithms are achieved by pitching two AI machines one against the other to identify the flaws and improve upon the earlier creations. This reiterative process creates continuously improving fakes until it reaches a stage where it becomes indistinguishable from the real video when it can get published.
The authentic data set used for learning may consist of hundreds or thousands of still photographs of a person’s face, so the algorithm has a wide selection of images showing the face from different angles and with different facial expressions to choose from.
Tomorrow if you receive a video call from your wife asking you to immediately transfer some money to some account, it is quite possible that the video call may actually be that of a fraudster who was earlier trying to fool you with a phishing e-mail or a voice call. The risk to the reliability of the Internet system is therefore extremely high.
Naturally, there is a thinking about how such deepfakes can be prevented. In US, it is said that a new law to criminalize deepfake is being considered.
In India we have so many anti nationals in the guise of journalists and activists that if we attempt to pass any law even to impose responsibilities on intermediaries to check the spread of fake news, immediately people rush to the Supreme Court alleging infringement of constitutional rights.
It is therefore time for us take a realistic assessment of the situation and ensure that irrespective of what the fake activists think, there is a need for a strong internet regulation that has to preserve the trust in the system. Otherwise the entire edifice of E Commerce and E Governance is in the danger of falling apart.
Presently, the amendments to Intermediary Guidelines under Section 79 of ITA 2000 is under consideration and it is time for the Government to take a tough stand on the intermediaries and make them responsible for fake news and liable for the consequences.