INTRODUCTION
Nowadays, AI has dominated the IT world with its numerous advantages as well as disadvantages. One of the mischief breeds of AI is the deepfake which has been leading the headlines for the past few days. Deepfake videos are created to make someone do such things or say something that they rather never do or say. According to the Oxford Dictionary, “deepfake is a video of a person in which their face and body are digitally altered so that they appeared to be someone else, typically used maliciously or to spread false information[1].” It uses Artificial Intelligence (AI) to make fake content by using some videos, images, or videos. Deepfake is derived from the combination of two concepts “deep learning” and “fake.” There are two technologies used to make a deepfake video, deep learning algorithms and Generative Adversarial Networks (GAN) or Variational Autoencoders (VA)[2]. Deep learning is a subgroup of machine learning that influences deep neural networks to synthesize facial expressions, lip movements, and other visual and audio elements. By using the technique of machine learning one can make a computer or a machine to imitate human behaviour. There are two neural networks in GAN, the generator, and the discriminator. The generator as the name suggests generates synthetic data that involves videos and images that look like actual footage of a person. The discriminator describes the difference between authentic and synthetic data by evaluating the generated content. This interaction between the generator and discriminator helps the generator to create content that is more difficult to discriminate from the real one. Deep learning and GAN lead to continual advancements in creating deepfake videos however, this technology has raised concerns due to the potential for misuse for creating fraudulent and misleading content.[3] A very recent example of this misuse of deepfake technology is the video of actress Rashmika Mandanna in which her face was morphed with the body of a social media influencer[4].
HOW TO IDENTIFY DEEPFAKE CONTENT
Deepfake videos can be used by using a deep learning algorithm. It identifies the unnatural blinking or an audio-visual mismatch. After training in deep learning algorithms, it can analyze the difference between real and deepfake content.[5]
There are unnatural blinking and facial movements in deepfake videos, by analyzing this one can identify the difference between a deepfake and a real video. Also, there is audio visual mismatch in deepfake content due to the lack of training in deep learning algorithms which makes deepfake content more identical.
[6]Another method to identify deepfake video is using the method of forensic bases detection. This method examines the audio and video data pattern by analysing the irregularities in shading and lighting of the video, and the geometric relation between the facial features.[7]
LEGAL PROVISIONS FOR DEEPFAKE IN INDIA
After the deepfake video of Rashmika Mandanna went viral, Delhi Police filed a complaint under section 465 IPC[8], section 469 IPC[9], section 66(c)[10], and section 66(e)[11] of the Information and Technology Act. Although there is no specific legal provision for deepfake in India, various provisions are implied to address the issue.
Information and Technology Act, 2000:
- Section 66: Section 66(c) of the IT Act describes the punishment for identity theft which is the imprisonment of a maximum of three years or a fine of one lakh rupees [12]and section 66(e) of the IT Act deals with the punishment for violation of privacy which is imprisonment for maximum three years or a fine of two lakh rupees or both.[13]
- Section 67: It deals with punishment for the person who publishes and transmits obscene material in electronic form which may extend to the imprisonment of 5 years or a fine up to 10 lakh rupees. Section 67A of the IT Act[14] deals with the punishment for publishing or transmitting sexually explicit acts, etc in electronic form, and section 67B [15]deals with the punishment for publishing or transmitting pornographic or sexually explicit acts of children in electronic form.
Indian Penal Code:
- Section 465 (Punishment for Forgery): A person committing forgery is punished either with imprisonment of a maximum of 2 years or a fine or both. [16]
- Section 469 (Forgery for harming reputation): A person coming forgery to harm the reputation of any party is being punished with imprisonment of a maximum of 3 years and also be liable for a fine.[17]
- Section 420 (Cheating)[18] and Section 468[19] (Forgery for cheating can also be considered while dealing with deepfake content.
Also, Deepfakes are considered an infringement of copyright under the Copyright Act of 1957. Section 52 of the Copyright Act consists of a comprehensive list of work not considered as infringement.[20] Deepfake is not a part of the list as it manipulates the existing content and hence is the infringement of copyright.
CONCLUSION
The creation of deepfake content has brought both exhilaration and trepidation. Although there are significant uses of deepfake in the entertainment and gaming industries, it has come up with various misuses. Deepfake can be used to spread malicious information, hate speech, and fake news about the government. Cyber terrorism, online fraud,
and identity theft by morphing are some of the mischiefs of deepfake technology.[21] Women can be the most targeted victims of deepfakes, by spreading non-consensual obscene or pornographic content, women can be exploited to a great extent.
After the matter of actress Rashmika Mandanna, the Indian government issued an advisory to all social media platforms to remove all the reported deepfake content within 36 hours failing to which will result in legal and judicial proceedings. After, the onus of protecting the privacy of the person lies with social media platforms which act as intermediaries under rule 3 of the Information Technology Intermediary Rule.[22]
Although there are many provisions to address deepfake issues, there is a need to update the Information and Technology Act and introduce new legislation to address the offences of deepfake technology. There must be support for advancing detection methods and promoting awareness about deepfakes and their misuse. Through the combination of innovative technologies, we can harness the potential of AI and safeguard against its misuse for a trustworthy digital future.
Author(s) Name: Mahi Jaiswal (ILS Law College)
References:
[1] Mika Westerlund, ‘The Emergence of Deepfake Technology: A Review’ (Technology Innovation Management Review, November 2019) <https://timreview.ca/article/1282> accessed on 27 November 2023
[2] Lutz Finger, ‘Overview of How to Make Deepfakes – It’s Scarily Simple’ (Forbes, 8 September 2022) <https://www.forbes.com/sites/lutzfinger/2022/09/08/overview-of-how-to-create-deepfakesits-scarily-simple/?sh=22b29c752bf1> accessed on 27 November 2023
[3] Lutz Finger, ‘Overview of How to Make Deepfakes – It’s Scarily Simple’ (Forbes, 8 September 2022) <https://www.forbes.com/sites/lutzfinger/2022/09/08/overview-of-how-to-create-deepfakesits-scarily-simple/?sh=22b29c752bf1> accessed on 27 November 2023
[4] ‘Government Issues Advisory to Platforms on Countering Deepfakes’, Mint, (Delhi, 7 November 2023)
[5] ‘How to Identify Deepfake Videos?’, Times of India, (7 November 2023)
[6] Ian Sample ‘What are Deepfakes – and How Can You Spot Them’ (The Guardian, 13 January 2013) <https://www.theguardian.com/technology/2020/jan/13/what-are-deepfakes-and-how-can-you-spot-them> accessed on 27 November
[7] Shubham Pandey and Gaurav Jadhav, ‘Emerging Technologies and Law: Legal Statues of Tackling Crimes Related to Deepfakes in India’ (SCC Online, 17 March 2023) <https://www.scconline.com/blog/post/2023/03/17/emerging-technologies-and-law-legal-status-of-tackling-crimes-relating-to-deepfakes-in-india/ > accessed 21 November 2023
[8] Indian Penal Code, s 465.
[9] Indian Penal Code, s 469.
[10] Information and Technology Act 2000, s 66C.
[11] Information and Technology Act 2000, s 66E.
[12] Information and Technology Act 2000, s 66C.
[13] Information and Technology Act 2000, s 66E.
[14] Information and Technology Act 2000, s 67A.
[15] Information and Technology Act 2000, s 26B.
[16] Indian Penal Code, s 465.
[17] Indian Penal Code, s 469.
[18] Indian Penal Code, s 420.
[19] Indian Penal Code, s 468.
[20] Copyright Act 1957, s 52.
[21] Shubham Pandey and Gaurav Jadhav, ‘Emerging Technologies and Law: Legal Statues of Tackling Crimes Related to Deepfakes in India’ (SCC Online, 17 March 2023)<https://www.scconline.com/blog/post/2023/03/17/emerging-technologies-and-law-legal-status-of-tackling-crimes-relating-to-deepfakes-in-india/ > accessed 21 November 2023
[22] ‘Government Issues Advisory to Platforms on Countering Deepfakes’, Mint, (Delhi, 7 November 2023).
