Abstract
Advancements in artificial intelligence have given rise to deepfakes, hyper-realistic manipulated media that pose significant challenges to legal frameworks worldwide. This research delves into the complex intersection of synthetic media and the legal landscape in India, focusing on the [1]Information Technology Act, 2000 (IT Act). We have undertook a comprehensive analysis of the [2]IT Act’s efficacy in addressing the multifaceted issues arising from deepfakes, considering privacy infringements, potential amendments, and the evolving nature of digital deception. This research paper explicitly talks about the possible criminal activities that can be committed by using Deepfakes and the possible section of the [3]IT Act that can be used to deal with these crimes. This research not only contributes to the understanding of the legal challenges posed by deepfakes but also proposes recommendations and suggestion that could be implemented in order to overcome these challenges, ensuring the [4]IT Act remains adaptive in an era dominated by synthetic media. As technology continues to evolve, this study aims to guide policymakers, legal practitioners, and scholars in navigating the dynamic landscape of digital deception within the Indian legal context.
Key Words
Deepfake, Information Technology Act, 2000, Artificial Intelligence, IT Act, Synthetic media.
Research Methodology
The Research methodology of this paper is to be descriptive, and the sources of research are secondary in nature that are published articles of the newspapers, blogs, and acts & notifications from the official website of the Ministry of Electronics & Information Technology.
Introduction
AI – Artificial intelligence is a new-age technology with a wide spectrum which also includes Deepfakes. Deepfakes is a synthetic media created using AI. It all started in 2017 when a reddit user with the user name Deepfakes and many other people on the same platform started sharing videos of celebrities with swapped faces. The feature of face swap which is available on many apps was meant to be for fun. However, it isn’t the same now.
The literal meaning of the term Deepfake is ‘deep learning’ which is the subset of machine learning methods, based on representation learning and artificial neural networks & fake is for something deceptive. These are hyper-realistic images, audios or videos produced by manipulating the facial appearances and audio through deep generative methods.
How deep fakes are made:
In order to create deepfake content, two crucial algorithms—the discriminator and generator—interplay intricately to form a Generative Adversarial Network (GAN). In this dynamic process, the discriminator’s job is to discern between real and artificially generated content, and the generator takes on the responsibility of creating synthetic digital content.
The key to this cooperative effort is an ongoing feedback loop in which the discriminator evaluates the fake content that the generator creates. The discriminator informs the generator when it has accurately determined whether the content is artificial or real. Through this iterative process, the generator can improve and hone subsequent deepfakes in response to feedback.
Within the GAN framework, the combination of these two algorithms creates a sophisticated learning environment. Using a variety of algorithms, GAN trains itself to recognize complex patterns on its own, which is necessary to generate realistic-looking fake images. The neural network, which is essential to this learning process, requires a lot of exposure to different datasets that include faces from different perspectives and in different lighting conditions. The neural network can identify the subtleties needed to create smooth and lifelike deepfake images thanks to this extensive dataset.
The intricate process of creating deepfakes is essentially captured by the cooperative dance between the discriminator and generator that is arranged within the GAN framework. These algorithms’ technical capabilities and adaptability in producing increasingly complex synthetic content are demonstrated by the complex learning process, which is fuelled by an abundance of diverse data.
While it can be difficult to spot deepfakes, some telltale signs include strange facial expressions, strange body movements, strange colouring, alignment issues, distortions when the video is slowed down or zoomed in, and uneven audio. It’s important to recognize, though, that these techniques might not always work, particularly when dealing with highly accurate deepfakes. The imperceptibility of such manipulations to the untrained eye emphasizes the necessity for sophisticated instruments and algorithms created especially to identify intricate manipulations in digital content. The constant cat and mouse game in the world of synthetic media is highlighted by the ongoing developments in both deepfake creation and detection. There are many apps and software are available on the internet which may not detect a very well-made deepfake. However, they can detect fewer complex videos or audio.
Review Literature
Synthetic media has become more well-known in the last several years and has become an essential tool for many different businesses. Well-known companies like Cadbury, Zomato, SunFest, and others have used deepfake technology to great effect in their marketing campaigns. These businesses have been able to create ads with a unique degree of personalization, catering to the individual tastes of their customers, thanks to this creative approach.
Notably, the application of deepfakes in advertising has gone beyond traditional bounds, allowing companies to create customized campaigns that connect with customers. This personalization goes so far as to include customers’ faces in ads, which makes it easier to create special messages during holidays. One thing that stands out in particular is how well-integrated the photographs of customers are with those of their favourite celebrities, which raises the relatability factor of these marketing campaigns. The usefulness of this marketing tactic lies on its ability to effectively target certain customer categories. Businesses that use deepfakes in their advertising have developed a more engaged and devoted customer base by creating a sense of relatability. This tactic raises the overall attractiveness of goods and services while also increasing brand visibility. The deliberate use of synthetic media in marketing efforts, which capitalizes on the nexus between technology innovation and consumer interaction, marks a paradigm change as the medium continues to develop.
In the film business, deepfake technology is becoming more and more useful, especially for post-production editing. Its use gives filmmakers an accurate tool to improve and correct situations, making it a more affordable option than reshooting whole sequences. Additionally, by producing realistic lip-syncing and enhancing the authenticity of actors’ performances, deepfake technology has the potential to revolutionize dubbing in motion pictures. This double use highlights how deepfakes may revolutionize film production by improving both artistic realism and cost-effectiveness.
Nevertheless, a number of ethical and privacy issues have surfaced as a result of deepfake technology’s widespread use. A critical analysis reveals that the disadvantages of this technology outweigh its benefits, despite its inherent risks and potential drawbacks. Given these factors, navigating the ethical terrain surrounding deepfake applications requires a prudent approach.
Misuse cases make social unrest and religious conflicts more likely, endangering the peace and stability of the world. Although some people might use deepfakes for sentimental reasons, the wider ramifications must be carefully considered, highlighting the necessity to balance the advantages and disadvantages of the technology. Some of the concerns are explained below:
Pornographic and explicit content – Concerningly, the amount of mass pornographic content with the faces of well-known actors or celebrities has increased as a result of the editing of both the visual and audio elements in videos. Because of how much content is readily available online, public figures are especially susceptible to the software’s careful analysis and replication features, which can produce incredibly lifelike deepfakes.
The disturbing trend has also led to an increase in online child pornography. [5]15,000 deepfake videos were posted online in September 2019, according to a study by the AI company Deep Trace. This is almost a twofold increase in just nine months. Out of all of these, an astounding 96% were porn, and 99% involved projecting female celebrities’ faces onto actors in adult content.
Financial Frauds – Many financial frauds that have caused enormous losses have been perpetrated in India in recent years. The use of voice clones or fake audios that are precisely altered to sound like their trusted person is how these crimes are carried out. [6]The head of a German energy company’s UK subsidiary transferred almost £200,000 into a Hungarian bank account in March of last year after receiving a phone call from a fraudster imitating the German CEO’s voice.
[7]83% of Indian victims reported experiencing financial loss, with 48% reporting losses of more than Rs 50,000, according to a McAfee report. [8]McAfee said that more than half (66%) of the Indian respondents said they would reply to a voicemail or voice note purporting to be from a friend or loved one in need of money.
Social and political unrest – Deepfake videos are becoming more and more common on the internet and social media, which is concerning, especially when they target prominent people in positions of authority. During the ongoing conflict, videos featuring the [9]presidents of Russia and Ukraine became prominent and spread misleading messages. While it was relatively easy to identify these videos as fakes due to their crude content, the potential impact on public perception is still concerning
[10]The US president is one of the political figures who is particularly vulnerable to deepfake manipulation. Artificial intelligence (AI) versions of President Biden’s voice were used in February to disseminate hateful messages, such as accusations against transgender women and untrue claims that men and women were being drafted to fight in Ukraine. These videos quickly amassed millions of views on the internet, demonstrating the power of false information of this kind.
Furthermore, the 2018 case involving [11]President Ali Bongo of Gabon serves as a reminder of the possible political consequences of deepfakes. His health-related rumours and worries resulted in allegations of a deepfake during a televised speech. Although the authenticity of the video is still under question, the incident shows how profoundly deepfake scandals can affect public opinion and political stability.). and A fabricated video of audio like this from a politician or well-known person can cause social and religious unrest in the country which might result in chaos and threat to the integrity of it. This could also lead in manipulating the elections by forwarding wrong videos of your opponent on the social media.
Manipulated audio or video material, particularly from prominent personalities, possesses the capability to provoke social and religious turmoil, resulting in disorder and jeopardies to a country’s sovereignty. Furthermore, this kind of content manipulation can happen during elections, where fake videos are shared on social media to discredit rival politicians. It can create misinformation and confusion over important issues. It could be also be used to harass and threat someone.
The challenge of verifying deepfake content exacerbates these problems. Differentiating between real and manipulated content is a complex and continuous task. Some may be easily identified due to their crude nature, while others require sophisticated software for in-depth analysis. The possible repercussions of unbridled deepfake spread highlight the urgent need for strong action to counter this escalating danger to social stability and public confidence.
Information Technology Act, 2000
In India, we do not have any specific legislation made to counter the problem created by AI or Deepfakes. [12]Information Technology Act of 2000 which extensively deals with cybercrime and e-commerce based on United Nation Modal law based on e-commerce of 1996. It is the primary act in India. This act applies not just to Indian citizens but also to people from other countries who might commit against the residents of India. the [13]IT Act of 2000, enable the usage of digital safes and guarantee the legitimacy of digital signatures. The digital record-keeping of files and attachments for government law enforcement commissions and approved private entities is included in the implementation. It also permits financial institutions to conduct transactions between parties, permits the central storage of user data, and gives banks permission, under the RBI Act of 1934, to openly record account holders’ folios in online ledgers for government authorities to view.
Cyber Appellate Tribunal:
The [14]Information Technology Act, 2000 provides for the establishment of the Cyber Appellate Tribunal. It was created under the [15]Section 48 of the Act. The CAT has the authority of a civil court and can hear cases involving cybercrime and data protection.
If an individual is dissatisfied with the decision of the Controller or Adjudicating Officer, he or she may file a complaint with the Cyber Appellate Tribunal (CAT), which has jurisdiction over the case.
To file an appeal, the individual must do so within 25 days of receiving the order, along with the applicable fees. Under certain conditions, the CAT may consider appeals filed after this period if it is satisfied with the reasons for the delay.
Furthermore, the CAT strives to resolve appeals within six months of their receipt.
Amendments:
In the [16]Prajwala case, the Indian Supreme Court in December 2018 ordered the government to develop and issue guidelines in two weeks to address content on online platforms that promotes child rape, rape, and pornography. The goal of this legal action was to quickly remove and deal with objectionable content. Further examining the effects of pornography on kids was a 2020 parliamentary report, which reflected broader societal concerns about the potential harm associated with explicit materials. When taken as a whole, these measures show how legislative and legal actions are being taken to address and lessen the harmful effects of inappropriate content—especially when it comes to minors—that is found online. This led to the amendments made in the Intermediary Guidelines and Digital Media Ethics Code Rules, 2018 of the IT acts 2000.
Under [17]Section 66 of the [18]Information Technology Act, 2000, which deals with computer-related offenses, two specific provisions, namely [19]Section 66C and [20]Section 66D, address aspects of cybercrimes which can invoked in the offense of identity theft or fraud.
Unauthorized Use of Electronic Identity: [21]Section 66C
According to Section 66C of the IT Act, using someone else’s password, electronic signature, or any other unique identifying feature dishonestly or fraudulently can result in legal repercussions.
Cheating through Personation with Computer or Communication Resources: [22]Section 66D
The act of personating someone in order to cheat by using computers or communication devices is the subject of [23]Section 66D. Legal action may be taken against anyone who uses electronic impersonation to commit fraudulent activities
If found guilty under the given provisions, the maximum sentence for imprisonment is three years. A fine of up to one lakh rupees may also be applied.
Penalties for Electronically Publishing or Transmitting Prohibited Content: [24]Section 67
Legal repercussions will follow anyone who disseminates, publishes, or permits the dissemination of pornographic, voyeuristic, or otherwise inappropriate content via electronic means to potentially corrupt and deprave those who come into contact with it. If found guilty for the first time, a fine of up to five lakh rupees and a maximum sentence of three years in prison are possible penalties. In the event of a second or subsequent conviction, the maximum sentence is five years in prison and a fine of ten lakh rupees.
Penalties for Publishing or Transmitting Sexually Explicit Content Under [25]Section 67A of the Act:
Legal repercussions may follow those who publish, transmit, or cause to be published or transmitted any electronic material that contains sexually explicit acts or conduct. If found guilty for the first time, you could face a fine of up to ten lakh rupees and a sentence of up to five years in prison. The penalties increase in accordance with a second or subsequent conviction.
Penalties for Publishing or Distributing Content in Sexually Explicit Acts That Feature Children: [26]Section 67B
This section deals with offenses pertaining to the publication or electronic transmission of content that shows children performing explicit sexual acts. The following activities are covered: producing, gathering, looking for, looking through, downloading, advertising, promoting, trading, or distributing such content. There are specific legal consequences for these offenses.
Cyberterrorism Penalties: [27]Section 67F
This section discusses cyber terrorism and details actions meant to jeopardize India’s security, unity, integrity, or sovereignty. These behaviours include introducing computer contaminants, trying to access computer resources without authorization, and preventing authorized personnel from entering the system. The severity of these offenses and their potential impact on national security are highlighted in the outline of the legal repercussions for cyberterrorism.
Under the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2022 of the Information Technology Act, 2000 these Crimes like misinformation, defamation and hate speech can also be prosecuted.
Along with [28]Information Technology Act, 2000 the following sections of [29]Indian Penal Code, 1860 can be invoked:
[32]Section 354C addresses voyeurism, which is defined as the non-consensual act of watching or capturing images of a woman in a private setting for sexual gratification. Offenders face up to three years in prison and a fine. [33]Section 354D refers to the crime of stalking. It criminalizes the act of following, contacting, or attempting to contact a woman despite her obvious disinterest or objection, causing her fear or distress. Stalking is also discussed in the context of online platforms.
[34]Section 509:
This section deals with acts intended to offend a woman’s modesty. It includes any word, gesture, or act intended to offend a woman’s modesty, making it a criminal offense. A violation of this section is punishable by imprisonment for a term of up to three years, a fine, or both.
The Indian Penal Code sections 354C, 354D, and 509 are also applicable to online sexual harassment. These provisions are broad enough to cover various types of harassment, including online harassment.
[35]Section 415: Cheating
This section talks about offenses of cheating. This section pertains to deepfakes, wherein individuals who engage in deceptive practices such as financial fraud or impersonation may face legal repercussions.
[36]Defamation: Section 499
Defamation charges under Section 499 are applicable when deepfakes are deliberately made or shared with the intent to damage someone’s reputation. Those who use deepfake technology in such a defamatory manner risk legal consequences.
Criminal Intimidation: [37]Section 503
When deepfakes are used to threaten or intimidate people, [38]Section 503 is triggered. This section emphasizes the seriousness of using deepfakes for coercive purposes and provides charges for those found guilty. When deepfakes are used for threatening or intimidating purposes, Section 506 comes into play. It penalizes offenders who use coercive deepfakes with legal consequences.
Intentional Insult: [39]Section 504
Intentional insult with the aim to cause a disturbance is covered by Section 504. This section may be used in the context of deepfakes when manipulations violate someone’s dignity and may have legal repercussions.
Sections 292 and 294 of the Indian Penal Code deal with the punishment for the sale and distribution of obscene material.
Suggestions
Deepfake technology has made India vulnerable to the spread of false information because of the country’s widespread lack of digital literacy. People tend to believe what they see. Significant funding and coordinated efforts are required to raise the country’s digital literacy levels in order to address this. Enforcing regulations is essential to managing online content in an efficient manner, preventing disinformation, and promoting a safe online environment. To ensure precise user identification and improve overall cybersecurity, multi-authentication for electronic signatures becomes essential.
In order to strengthen defenses against new threats in the digital landscape, strategic investments in cutting-edge technologies are essential. Investing money to create advanced instruments will greatly improve our ability to address changing problems. Furthermore, it is impossible to overestimate the significance of routine reviews and modifications of current regulations. The implementation of an iterative approach guarantees the resilience, responsiveness, and alignment of legal and technological frameworks with the ever-changing demands of the modern world. When taken as a whole, these actions provide a thorough plan of action to protect against the threats presented by false information and new technology, fostering a safe and knowledgeable digital community.
Conclusion
When dealing with the modern problems brought on by the quickly changing technology environment, it is essential to acknowledge the dynamic character of problems that change on a daily basis and to put equally flexible solutions into practice. Due to its profound benefits and potential drawbacks, artificial intelligence (AI) demands strict regulation in order to reduce the likelihood of criminal activity being enabled by this technology. Sites and platforms that encourage the misuse of deepfake technology should be prohibited, as this technology presents a serious threat.
The sheer size of social media and online platforms highlights the shortcomings of having a single, inflexible legal framework. Rather, it is imperative to adopt a proactive and ongoing strategy that involves frequent reviews and updates. This guarantees that the legal framework will continue to be flexible and responsive to new issues. By recognizing the need for ongoing monitoring and creative fixes, we can create a regulatory framework that can successfully handle the complex and ever-changing problems brought on by the use of contemporary technologies.
NAME: KSHITIJA SHIVANKAR
COLLEGE: GOVERNMENT LAW COLLEGE, MUMBAI.
[1] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[2] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[3] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[4] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[5] Doctored Sunak Picture is Just Latest in String of Political Deepfakes, THE GUARDIAN (August 3, 2023), https://www.theguardian.com/technology/2023/aug/03/doctored-sunak-picture-is-just-latest-in-string-of-political-deepfakes.
[6]Dustin Volz, Fraudsters Use AI to Mimic CEO’s Voice in Unusual Cybercrime Case, WALL STREET JOURNAL (August 26, 2019), https://www.wsj.com/articles/fraudsters-use-ai-to-mimic-ceos-voice-in-unusual-cybercrime-case-11567157402.
[7] Connected Family Study 2022 India, MCAFEE (May 12), https://www.mcafee.com/content/dam/consumer/en-in/docs/reports/rp-connected-family-study-2022-india.pdf.
[8]Author Not Specified, about 83% Indians have Lost Money in AI Voice Scams, TIMES OF INDIA (May 1, 2023), http://timesofindia.indiatimes.com/articleshow/99914367.cms.
[9] Jane Wakefield, Deepfake Presidents Used in Russia-Ukraine War, BBC (March 16, 2022), https://www.bbc.com/news/technology-60780142.
[10] Factcheck: Biden Did Not Make Transphobic Remarks, REUTERS (May 12, 2023), https://www.reuters.com/article/factcheck-biden-transphobic-remarks-idUSL1N34Q1IW.
[11] Michael Birnbaum, How a Sick President and a Suspect Video Helped Sparked an Attempted Coup in Gabon, WASHINGTON POST (Feb. 13, 2020), https://www.washingtonpost.com/politics/2020/02/13/how-sick-president-suspect-video-helped-sparked-an-attempted-coup-gabon/.
[12] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[13] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[14] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[15] Information Technology Act, 2000, No. 21, 2000, India Code (2000),§48, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[16] Information Technology Rules, 2021, WIKIPEDIA (last modified October 25, 2023), https://en.wikipedia.org/wiki/Information_Technology_Rules,_2021.
[17] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §66, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[18] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[19] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §66C, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[20] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §66D, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[21] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §66C, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[22] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §66D, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[23] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §66D, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[24] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §67, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[25] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §67A, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[26] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §67B, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[27] Information Technology Act, 2000, No. 21 of 2000, INDIA CODE (2000), §67F, https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[28] Information Technology Act, 2000, No. 21, 2000, India Code (2000), https://www.indiacode.nic.in/bitstream/123456789/13116/1/it_act_2000_updated.pdf.
[29] The Indian Penal Code, India Code (1860), https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[30] The Indian Penal Code, India Code (1860) §354C, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[31] The Indian Penal Code, India Code (1860) §354D, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[32] The Indian Penal Code, India Code (1860) §354C, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[33] The Indian Penal Code, India Code (1860) §354D, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[34] The Indian Penal Code, India Code (1860) §509, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[35] The Indian Penal Code, India Code (1860) §415, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[36] The Indian Penal Code, India Code (1860) §499, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[37] The Indian Penal Code, India Code (1860) §503, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[38] The Indian Penal Code, India Code (1860) § 506, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
[39] The Indian Penal Code, India Code (1860) §504, https://www.indiacode.nic.in/bitstream/123456789/2263/1/aA1860-45.pdf.
Hey it’s Sage, here is the free AI tool that automatically writes copy to drive traffic and calls
Go to:
http://www.autoSEOcontent.com
Hope it helps