Compromised identities

On July 9, a man in Kerala, despite exercising aconsiderable degree of caution, lost Rs 40,000 to AI fraudster(s) who duped himthrough deepfake technology. Deepfake is a hybrid term denoting the combinationof ‘deep learning’ — a subset of machine learning — and ‘fake’. It refersto the use of artificial intelligence (AI) and deeplearning techniques to create or manipulate audio and video content in a waythat makes it appear as though someone is saying or doing something they neveractually did. Though it has numerous positive applications in the present age,deepfake has earned notoriety for being misused for the purpose of carrying outfrauds. Fraudsters usually train deep learning algorithms like convolutionalneural network (CNN) or a generative adversarial network (GAN) on a vast amountof data pertaining to the individual target they have set. For example, in thecase of the Kerala man, the fraudster(s) impersonated his former colleague withwhom he had worked for four decades. They chatted with the frauded man onpersonal whereabouts and other information, including about his daughter. Whenrequested for money on voice call, citing health emergency, the target wascautious enough to seek cross verification through video call. That, however,was not enough. The fraudster(s) had already engineered a video call whereinthe deepfake version of his former colleague talked to him over video callwithout arousing any suspicion. This case, though not first of its type, forcesone to reflect on two crucial things. First, can humans still unequivocallytrust their ears and eyes for what they hear or see? Second, can deepfakestill be ignored as being a technology in its infancy? Unfortunately, theanswers to both these questions are in negative. Humans cannot necessarilybelieve what they see or hear over the phone, and deepfake has quickly grown tobecome more than an infant! The case of Kerala man is still an incident ofbearable monetary fraud. In the past, deepfake technology was misused toerroneously showcase that a nation head (Ukrainian President VolodymirZelenskyy) was willing to surrender in a war that affects the entire world.Quite recently, the face of an Australian woman was morphed onto a pornographicvideo content, and widely circulated. Attempts were also made to deepfakenumerous famous personalities including Jennifer Lawrence, ArnoldSchwarzenegger, and Mark Zuckerberg, among others. The high-profile cases aremost likely to be identified and refuted, but those relating to monetary fraudand privacy breach on common citizens can pass initial scrutiny, letting thedamage be already done before the scam is detected. This necessitates theformulation of specific laws around curbing deepfakes. While selected countrieshave made ground-breaking progress in this direction, most of the countries,including India, have displayed lack of proactivity and technological capacityin this direction. It may be noted that negative applications of deepfaketechnology extend far beyond monetary frauds and privacy breaches. It can beused to spread propaganda, misinformation and provocative content —particularly during election times. In a world that is increasingly gettingpolarised, deepfakes hold immense potential to create a wedge in society. Inwake of these considerations, formulation of a specific legislation assumesparamount importance. The complex and evolving nature of deepfake technologymakes it too big a fish to be caught in the net of certain generic IT ruleprovisions. Until the government comes up with effective laws against deepfake,or even after that, people need to exercise extreme caution. It has to beadmitted that what we see and hear in this age may be a complete lie. In fact,we need to stick to the Zero Trust approach which assumes that our networks arealready breached, computers already compromised, and all users are potentialrisks. The approach forbids trust without verification. The time has alreadypassed when citizens could seek solace in the fact that emerging technologiesare in their infancy. They are maturing fast, setting the world in turbulence!People need to have their feet stable before they are forced to fall.