Unmasking Deep Fakes: Perception at stake
If you are a Tom Cruise fan, you must know Miles Fisher. The TikTok star, famous as ‘Tom Cruise Guy’. He mirrors the actor wholly with the help of deep fakes and creates funny content. Miles already had the mannerisms and body structure, all he required was the actor's facial features to become Tom Cruise, and he got it. It's an attestation to the power of AI-generated deep fakes in terms of their ability to create a person out of another person.
In the past few years, the use of Hyper-Realistic Deep forgeries (Deep fakes) has skyrocketed. Post 2020, a handful of directors made use of it extensively to translate their films into different languages and de-age veteran actors & actresses. Recently visual effects artists used De-ageing software on Harrison Ford to reduce his age to 40 years in the latest part of Indiana Jones.
But first, let’s learn how deep fakes are created; Leonardo da Vinci, to his Mona Lisa (deep fakes), is GANs. Generative Adversarial Networks are a computing process in which two AI, Generator and Discriminator, work in a cycle.
'Generator', as the name suggests, generates deep fake images. Collecting multiple photos of a person covering the full face. It produces an initial deep fake and then sends it to the ‘Discriminator’ (Detective), which detects errors in the image and sends it back to the generator. This cycle continues until the best deep fake is generated.
One thing to remember about deep fakes is you need many photos from different angles to create one. That’s why celebrities are the first ones to get attacked by it.
The tech is cool as far as innovation is concerned. It pushes the boundaries of imagination and stretches the horizon of the human intellect, but most people are using it to create chaos in the world. These monsters behind the masks want to create a colosseum on the web and ignite a fire, and deep fakes are their Molotov cocktails.
The face of America's richest man was recently used to promote a Bit-coin trading platform called 'BitVex'. Elon Musk was scammed! A look-alike of him was giving an interview about a trading platform, and his face got swapped, along with his voice, and boom, another copy of Mr. Musk was produced.
But this isn’t just limited to promoting something; according to Sensity, a company formed to detect deep fakes published a report in 2019 that 96% of deep fakes are used on women without consent, to create pornographic content.
The process goes like this - if a woman has multiple photos of herself on social media accounts, the deep fake software collects them and then puts them on the face of an adult film actress. And soon, it reaches everyone's phone gallery. Later, when the girl finds out, her world is brought to a standstill, and distress digs a deep hole into her unconsciousness. Often, suicidal thoughts penetrate and take the girl's life story into some undescribed location.
The only country which has taken some tangible action against the deep fakes is China. They call it 'deep synthesis technologies' and regulate the content by marking every deep fake content as 'modified'. Moreover, the European Union is also catching up. They're working on putting a label on the deep fake content.
The meat of all deepfake progress is; it made people more skeptical of online content, leading to increased awareness and critical thinking. However, there is a risk of genuine footage being mistaken as fake, which could have negative consequences. As this emerging technology is enhancing at every step, it's testing human intelligence to judge what is white & black.
