Rashmika video reveals how scary AI deepfakery can be

Actor Rashmika Mandanna became the unfortunate target of a deepfake video that went viral recently. Tomorrow it could be one of us, our children, or even grandchildren. All it takes is a photo or video from a social media profile and an app that uses artificial intelligence (AI). Identity theft is possibly the most dreaded thing for a person. Just imagine the shock that comes with realising that your image can be manipulated to do things you would never do. Rashmika’s impassioned plea on social media drew attention to this new menace. "We need to address this as a community and with urgency before more of us are affected by such identity theft," she implored, pointing out that had such a terrible thing happened to her when she was in school, she would not have known how to handle it.In Rashmika’s case, British influencer Zara Patel, who was in the original video, was quick to deny involvement and issue a warning: “Not everything on the internet is real…” That’s just it! Who can tell what’s fake, deepfake or deepest fake?Soon after this incident, a morphed video of superstar Katrina Kaif began doing the rounds. Purported to be a leaked clip from her forthcoming movie, it featured a fight sequence between two women dressed in towels. Movie stars are far more vulnerable to such demeaning, dehumanising acts which are designed to titillate and mislead their enormous fan base. But it’s not just celebrities — it’s as easy to digitally undress an ex-girlfriend and circulate the image or video to the world as revenge. A 2019 report by Sensity, a company that monitors deepfakes, found that 96% of deepfakes were non-consensual sexual deepfakes, and of those, 99% featured women.However, there seems to be very little experts can do about the menace. This is just the tip of the iceberg, warn global techies, drawing attention to how easy it is to be manipulated as AI-generated imagery gets disturbingly sophisticated.Generative AI has another dark side — chatbots that make things up. Experts call them ‘hallucinations’. Worse still, there is no accurate way of predicting how often these things ‘hallucinate’. A New York Times report cited research done by a new start-up called Vectara which estimated that chatbots invent information at least 3% of the time — and as high as 27%. This is far from reassuring news, as we rely increasingly on these AI tools. ShowBlurbsIf that’s not enough, audio deepfakes which use cloning software are on the rise. A few years ago, the head of a UK Energy company was fooled into sending cash to his German CEO on the basis of a “call’’ in his voice and accent. According to a report by McAfee, about half (47%) of Indian adults have experienced or known someone who has experienced some kind of AI voice scam. What would a parent do if they received a distress call in their child’s voice? Would the panic-stricken parent have either the time or resources to crosscheck the veracity of the voice before wiring funds? Deception at this level can seriously impact the lives of common citizens, not just high-profile celebrities.Deepfake creators are ten paces ahead of those tracking them. So far, we have not come up with watertight solutions or even basic protection of our dignity and autonomy. Author Salman Rushdie nailed it when he said: “So, freedom is under pressure from the left as well as the right, the young as well as the old. This is something new, made more complicated by our new tools of communication, the internet, on which well-designed pages of malevolent lies sit side by side with the truth, and it is difficult for many people to tell which is which."Some prominent Hollywood stars like Tom Hanks and Scarlett Johansson have raised an alarm about the use of AI-generated voices and likenesses in advertisements. The latter has even taken legal action against an app. But what do people outside these charmed circles do to retain control over their identity, stay safe from potential scammers or even sift the truth from the deepfake news? In the Ukraine-Russia conflict as well as the Israel-Hamas one, the use of deepfakes raised concerns.The Mandanna video is a wake-up call that something needs to be done to keep up with the complex, rapidly evolving world of artificial intelligence. Social media platforms especially have to do more to control the spread of deepfakes. Without proper regulation, it will be difficult to harness the benefits of AI while protecting the vulnerable from its risks. Till that happens, don’t take anything at face value.