BY ANTONY MUTUNGA
As the Israel-Hamas crisis intensifies, millions of citizens, especially ordinary individuals, are now active on social media platforms – they are directly affected and thus, in need of aid from the rest of the world, with many calling for donations through different channels.
However, it is disheartening that most of the “soft requests” are created by hackers and cyber-criminals taking advantage of the war to enrich themselves at the expense of the true victims of the crisis.
The use of deepfake technology, which entails manipulation of facial appearance through deep generative methods, has become popular and quite accurate at manipulating users. We live in dangerous times as some techies engage powerful techniques like machine learning and artificial intelligence (AI) to not only manipulate, but also generate visual and audio content that can easily hoodwink online users.
The advancement of AI has seen deepfake technology evolve big time. Able to mimic one’s appearance and voice with accuracy, this technology in the hands of the masses, millions have already unknowingly fallen victim to it and worse off, some have even lost their earnings or donations thanks to deepfake scams. Social media is rapidly gaining ground with different players such as Facebook, Instagram and Tiktok vying for a piece of the pie. The bad news is that cyber criminals are now using such platforms to target millions of clueless online users.
The emergence of deepfakes adds to the arsenal of hackers. Now vicious hackers use deepfakes to imitate not only celebrities, but also people in need of assistance. A number of Hollywood celebrities and famous figures have seen their identity used without their knowledge for advertisements, and sometimes for distortion. Actors such as Tom Hanks and Robert Pattinson, have been victims to such scams.
In 2022, a TikTok account claiming to be set up by Robert Pattinson amassed over 600,000 followers a few weeks after being set up. If it wasn’t for a noticeable error in copying the unique shape of Pattison’s ear, no one would have noticed it was a deepfake.
On the other hand, a deepfake of Tom Hanks was used in advertisement that was promoting a dental plan, leading the actor to come out and warn his millions of followers that it was a scam.
The main challenge is hackers have gone a step further, and are now pouncing on donations that are for those in need such as victims of the ongoing conflicts across the world. Through deepfakes, cyber-criminals are able to mimic officials of charity organizations and leaders of economies in conflicts through social media, and mix them with genuine charity material to lure people to make donations through false channels. Deepfakes of President Volodymyr Zelenskyy of Ukraine in 2022, asking for donations through other channels than the official ones, still ring a bell.
To date, the number of deepfakes has gone up with data from the World Economic Forum indicating that deepfakes continue to grow at an alarming annual rate of 900%. In an attempt to tackle the evolving threat, social media platforms have moved to ensure their systems are able to detect and verify deepfakes using AI algorithms. However, this is not a simple task as hackers are always ahead of the game, and ever evolving.
Also, with not every deepfake used negatively, platforms need to balance the right of users to express themselves freely and creatively with the responsibility to prevent and mitigate the harms caused by deceptive deepfakes.
This has seen social media platforms decide to invest in research and innovation to improve their systems using AI techniques that include deep learning, computer vision, natural language processing, and audio analysis, in an effort to fight the deepfakes. In fact, a number of platforms have introduced deepfakes to their systems in order to improve their accuracy to detect deepfakes.
Facebook is one pioneer social media platform that has recognized the increasing threat and moved to teach its system and AI software to be able to spot manipulating videos by first introducing a database of 100,000 deepfakes in 2020, of which it was able to detect 65%.
Following in its steps, social media platforms have made it a routine to keep testing its systems to keep up with deepfakes.
It is also the responsibility of platforms to educate and create awareness on the dangers of deepfakes, in order to help users identify the misleading medias. Social media platforms can further introduce tools and features to help users identify, report, flag, or challenge deepfake content on their platforms. It could start by providing visual, auditory cues, or indicators to signal that a content is a deepfake.
Every user should be aware that hyper-realistic deepfakes are now a reality and that there is a need for caution. As we head into the festive season, a period of joy when consumers will be immersed in the metaverse, there is a need to tread lightly (online) as deepfakes are here to stay and we are headed to their domain.
