Deepfakes

DEEPFAKES

Deepfakes use a form of artificial intelligence called deep learning to make visual and audio content of fake events, hence the name deepfake.

The origin of the word “deepfake” can be traced back to 2017 when a Reddit user, with the username “deep fakes”, posted explicit videos of celebrities.

Deep Fakes are created by deep learning models, which use neural networks to manipulate images and videos.

A model such as this “analyses” video footage until it is able algorithmically to transpose the ‘skin’ of one human face onto the movements of another.

The usage of a technique called Generative Adversarial Networks (GAN), which employs two AI algorithms — one that creates false material and the other that assesses the system’s efforts, allowing it to improve has resulted in more accurate deep fakes.

Uses of Deepfake Technology

• Health and medical care: Deep generative models can be used to improve healthcare. These models—generative adversarial networks, or GANs—are especially helpful in overcoming data scarcity caused by privacy issues or the rarity of specific medical diseases. They may provide lifelike medical images, including brain MRI scans, which improve the AI’s learning process and enable more accurate diagnosis and therapy

• History and Art: We can better connect with and comprehend the past by using deepfake technology to reconstruct it. It can provide us the chance to encounter things that predate us and that we otherwise wouldn’t be able to understand.

• Personalization: A strong argument may be made for the appropriate commercial application of deepfakes, which would enable large-scale personalization. While personalization is a wonderful thing, many firms nowadays struggle to use it on a large scale. Thus, the business use case for deepfake technology is video personalization.

Concers Regarding Deepfake

• Deepfakes can be used for a variety of nefarious activities, including:

• Disseminating false information and propaganda;

• Swaying public opinion and elections;

• Blackmailing and extorting people or groups;

• Destroying the credibility and reputation of public

figures, politicians, activists, and journalists;

• Producing non-consensual and revenge pornography.

• Deepfakes can undermine the rule of law and human rights, as well as erode public confidence in democratic institutions, the media, and democracy itself.

• People’s privacy, dignity and reputation can be violated by deepfake technology, which can also be detrimental to victims’ mental health and general wellbeing particularly women, who are frequently the objects of such nefarious manipulation.

Global Initiatives to tackle Deepfake

• 28 major countries, including the US, China and India came to an agreement at the world’s first AI safety summit in 2023 about the necessity of taking global action to address the possible threats associated with AI.

• The summit’s Bletchley Park Declaration recognized the dangers of purposeful abuse and the loss of control over artificial intelligence (AI) technologies.

India and Deepfake regulations

• Deepfake technology is not specifically prohibited or regulated by laws or regulations in India.

• A worldwide framework for the development of “ethical” AI tools has been demanded by India.

• Deepfakes can be subject to legal action for some of their activities, including as releasing explicit content and defamation, under sections 67 and 67A of the Information Technology Act (2000).

• The Indian Penal Code (1860), Section 500, stipulates the penalties for defamation.

• Digital Personal data Protection Act 2023: According to the definition provided by the Digital Personal Data Protection Act of 2023, pictures and photos are considered sensitive personal data that can be used to identify a specific individual. Thus, deepfakes constitute a breach of personal information and an infringement on an individual’s right to privacy.

Way Forward

• Creating and putting into effect extensive legislation and rules that, while maintaining the right to free speech and expression, particularly target the production and distribution of deepfakes.