DEEPFAKES

1. Context

2. What are Deepfakes

• Deepfake is a type of synthetic media in which a person in an already-existing video or image is replaced with another person. It manipulates the audio/video, which has the propensity to the device, using machine learning and artificial intelligence.
• Due to the ease with which bogus news, celebrity pornographic content, etc. get shared online, it has drawn attention.
• It makes a fake version of original or real audio-visual content by superimposing a new audio or image over an existing media file.
• In September 2019, the AI company Deeptrance discovered 15,000 deep fakes videos online-nearly tripling in just nine months. A starting 96% of them were pornographic, and 99% of them matched the faces of famous women to porn actors.
• Deepfakes can be used to damage reputation, fabricate evidence, defraud the public, and undermine trust in democratic institutions.
• All this can be achieved with fewer resources, with scale and speed, and even microtargeted to galvanize support.

• Deepfake content is created by using two competing AI algorithms- one is called the generator and the other is called the discriminator.
• The discriminator is tasked with determining if the fake multimedia content produced by the generator is real and manufactured.
• A generative adversarial network is created when the generator and discriminator work together (GAN). Every time the discriminator correctly recognizes the content as being fake, it gives the generator important insights into how to make the next deep fakes better.
• The first step in establishing a GAN is to identify the desired output and create a training dataset for the generator.
• Video clips can be supplied to the discriminator after the generator starts producing output at a level that is acceptable.

• The first case of malicious use of deep fake was detected in pornography. According to sensity.ai, 96% of deepfakes are pornographic videos, with over 135 million views on pornographic websites alone. Deepfake pornography exclusively targets women.
• Pornographic deepfakes can threaten, intimidate, and inflict psychological harm. It reduces women to sexual objects causing emotional distress, and in some cases, leading to financial loss and collateral consequences like job loss.
• Deepfake could act as a powerful tool by a malicious nation-state to undermine public safety and create uncertainty and chaos in the target country. Deepfake can undermine trust in institutions and diplomacy.

• Deepfake causes financial fraud, which poses problems for the entire financial system.
• In the era of the threat of fake news, it also poses a threat to the security of cyber systems and the validity of online registration.
• Deepfakes in phishing efforts would make it more challenging for people to recognize a hoax.
• In any nation, deep fakes can be used to sabotage democratic procedures like elections.
• The potential for harm to people, organizations, and societies is enormous since it can be used to generate phony pornographic videos and make politicians appear to say things they did not.
• Any genuine evidence of a crime can be easily discounted as false because the public is so distrustful due to the prevalence of deep fakes.
• Fake movies are likely to become more popular outside the world of celebrities as new technology enables unskilled people to create deep fakes with just a few images. This will feed the growth of revenge porn.
• The use of fake identities and impostor frauds in cybercrime is rising.

6. What is the Solution?

• Media literacy efforts must be enhanced to cultivate a discerning public. Media literacy for consumers is the most effective tool to combat disinformation and deep fakes.
• We also need meaningful regulations with a collaborative discussion with the technology industry, Civil society, and policymakers to develop legislative solutions to disincentivize the creation and distribution of malicious deepfakes.
• Social media platforms are taking cognizance of the deepfake issue, and almost all of them have some policy or acceptable terms of use for deepfakes.
• We also need easy-to-use and accessible technology solutions to detect deepfakes, authenticate media, and amplify authoritative sources.