I. ABSTRACT
At present we live in a world where it is difficult to make out what is real and what is fake, while manipulation and deceiving has been an old age art practiced over centuries, the difference between the real and surreal is getting to a point being “NIL”. Deepfake, also known as synthetic media uses Artificial Intelligence to manipulate images, videos, audios in order to deceive the average person and make them believe what they are seeing is the truth. Deepfake applies Generative Adversarial Networks (GANs) to create digital illusion and manipulate reality. First originated in the year 2017, when a moderator on the Reddit online platform started uploading photos in which he swapped the faces of adult film performers. Presently deepfakes are either used for malicious activities or for spreading misinformation creating problems such as violation of privacy, human rights, data breach and many more.
II. INTRODUCTION
Deepfakes are video, audio and images that are edited and manipulated with the help of artificial intelligence. The first deepfake video was first generated in the year 2017, when a Reddit Moderator created a subreddit (small community) known as “deepfake” where the moderator swapped the faces of celebrities with people involved in adult content.[i]
The prevalence of deepfake seems to be increasing with time and is threatening to assume Frankenstein monster like proportions. These days, it has become very difficult to distinguish between an original video and a duplicate one. The recent newspaper reports of Microsoft predicting large scale disruptions in the upcoming general elections[ii] and previous instances of actor Anil Kapoor’s voice impersonating and the recent case of pop-singer Taylor Swift’s deepfakes are few of the many instances of deepfakes issue.
The answer to combat and overcome this harmful use lies in Artificial Intelligence based technology itself. Similar to dealing with a virus operating in the system of a computer where an antivirus program is required to detect, control and eliminate the threat, strong counter measures based on AI is the need of the hour.
Intersection of Deepfake and Intellectual Property Rights
Deepfake technology relies mostly on copyrighted materials for generating content. The technology uses algorithms that digest these vast amount of data/materials which includes real images, voices, and videos. The algorithm then analyze the data sets and create content that replicates the person’s voice, likeness, facial feature, movements with such accuracy that tricks people into believing that the content involving the individual is genuine, even though such acts didn’t happen in reality.[iii]
By using copyrighted works without permission, violating personality rights brand image of celebrities by mimicking their voice, likeness, images without permission, deepfakes pose a significant threat to the IP rights of individuals.
Recently, a group of artists signed an open letter from the Artist Rights Alliance calling out AI companies and developers, the letter highlighted the predatory use of AI technology in the form of “deepfakes”, “voice cloning” and unauthorized use of musical works to train AI models.
A part of the letter sent read as follows: "Some of the biggest and most powerful companies are, without permission, using our (musicians & artists) work to train AI models. These efforts explicitly target at the substation of human artists with huge quantities of AI-created sounds and images thereby diminishing the royalty pools that are paid to artists. This could be disastrous for many musicians, artists and songwriters who are trying to make ends meet”[iv]
While Artists and Musicians are fighting against the possibility of AI generated voices taking over music industries. Actor and Actresses like Rasmika Mandanna, Anil Kapoor, Alia Bhatt and many others too have been the victim of this misused technology.
Last year, a deepfake video of Actress Rahmika Mandanna created by a fan went viral on various Social Media platform, where the actress’s face was morphed on to the face of a British Influencer Zara Patel.[v] This incident stirred up a lot of discussion amongst the people on the potential misuse of the technology urging the government to regulate the technology.
Similarly in the very recent case of Anil Kapoor v. Simply Life India & Ors[vi], Actor Anil Kapoor (Plaintiff) filed a suit against the Defendants for utilizing Deepfake Technology and Artificial Intelligence and representing Plaintiff as actresses in a derogatory and offensive manner, using Plaintiff’s generative Artificial Intelligence as a Disney Character, Morphing Plaintiff’s face and seeking payments from customers etc.,[vii]The Hon’ble Delhi High Court took cognizance of the case and passed an ex-parte injunction order against the Defendant restraining them from using Plaintiff’s “name, likeness, image, voice, personality or any other aspects of his persona to create any merchandise, ringtones, ring back tones, or in any other manner misuse the said attributes using technological tools such as Artificial Intelligence, Machine Learning, deepfakes, face morphing, GIFs either for monetary gains or otherwise to create any videos, photographs, etc., for commercial purposes, so as to result in violation of the Plaintiff’s rights” Despite the issues faced by deepfakes and AI manipulation, technological breakthroughs are assisting in the invention of software, tools and algorithms to detect and curb such fraudulent acts, therefore protecting individual’s rights and preserving integrity of digital information.
Combating Deepfake using Technology
When it comes to tackling the issue of deepfakes, technology plays a crucial role. The inventions made by companies and researchers serve as a testament of progress against the fight of deepfakes. Some of the technologies that are currently being used to detect deepfakes are as follows:
Intel’s Real-Time Deepfake Detector
Intel had invented a real time deepfake detector called Fakecatcher. This technology detects fake videos with an accuracy of 96%. The technology checks for “blood flow” in the video pixel that identifies as humans.[viii]
Sentinel The technology allows individuals to upload digital media like pictures on to their website or API, to check for AI manipulation. Sentinel uses powerful artificial intelligence algorithms to analyze uploaded media and determine whether it has been manipulated or not and later provides a visual representation of the manipulated media.[ix]
Deepware AI
The technology is trained on authorized videos collected from YouTube, 4Chan and Celeb-DF platforms to remain up to date with new online trends. The detector’s wide collection of video content helps it to identify synthetic media precisely.[x]
Sensity AI
The technology is trained to find GAN framework to detect deepfakes. The technology utilizes deep learning and computer vision to detect and keep a track on AI generated synthetic media.[xi]
Microsoft’s video Authenticator Tool
Microsoft invented the Authenticator Tool to identify deepfake videos and photos, As the media plays the application provides a “confidence score or percentage score” in each time frame to detect deepfakes which are missed by the human eye.[xii]
III. TECHNOLOGIES FOR WHICH PATENT APPLICATIONS HAVE BEEN FILED IN INDIA TO CURB THE DEEPFAKE MENACE.
To curb the deepfake hazard, India too has embarked the journey of technological innovation. Here are some patent application fields by inventors for detection of deepfakes.
Application number | Earliest patent priority date | Title | Inventors |
IN202121026931 | (IN202121026931)2021-06-16 | (IN202121026931)System and method for detecting deepfake image and video | Suratkar Shraddha SunilRanjan PranjalPatil SarveshParekh AnkitMalolan BadhrinarayanSingh SakshamKazi Faruk |
IN202121026913 | (IN202121026913)2021-06-16 | (IN202121026913)System and method for detecting deepfake image and video | Suratkar Shraddha SunilRanjan PranjalPatil SarveshParekh AnkitMalolan BadhrinarayanSingh SakshamKazi Faruk |
IN202341085040 | (IN202341085040)2023-12-13 | (IN202341085040)Detecting deepfake videos on social media using machine learning | Florence Bharathi RPandya ApurvakumarBhandari SmitaAsati MukeshBarde SnehlataTembhurne AmitKumar ParveenMerlin Sasikala J ENagpal Pooja |
IN202321084571 | (IN202321084571)2023-12-11 | (IN202321084571)Machine learning-based system to detect deepfake videos on social media | Kuri Moushmee MilindPathak Ganesh RMakubhai Shahin ShoukatRamteke Nilima SurendraShendkar Bhagyashree DineshNandimath Jyoti NavinKamble Pratik Bibhishan |
IN202311085440 | (IN202311085440)2023-12-14 | (IN202311085440)Machine learning-based system to detect deepfake videos on social media | Sharma M KYadav HarendraKumar TarunSharma AjendraChaudhary SadhanaSingh DhanpalSrivastava MayankDoshi DhairyaAgrawal Pratik KDhuri Kanchan GaneshSingh PoonamShandilya KaushikKumar MukeshSanwal TanushreeKumar JaideepGupta Mohit |
IN202441011947 | (IN202441011947)2024-02-20 | (IN202441011947)Patented techniques for combating deepfake content in social media through ensemble machine learning | Nellore Manoj KumarSingh AjaiJagadish HemaKaur SonampreetAgrawal GauravKrishnan PrabhakarChoudhary MeenakshiSankar Siva B |
IN202121052989 | (IN202121052989)2021-11-18 | (IN202121052989)Automatic audio authentication software product for detecting deepfake voice and all types of audio tampering such as phase and spectrum discontinuity, pauses, dc shift, copy-paste, and background noise change | Dinkar Apte Shaila |
IN202241021964 | (IN202241021964)2022-04-13 | (IN202241021964)System and method for body parameters-sensitive facial transfer in an online fashion retail environment | Sukhwani Mohak Kumar |
IN202441017014 | (IN202441017014)2024-03-09 | (IN202441017014)Design and development of the deepfake detection framework | Bagath Basha CAmutharaj JAnuradha TRajalingam BRethinavelan SCharanarur PanemJose AnandNafizaKamarajapandian PKrishnaprasanna R |
IN202441020271 | (IN202441020271)2024-03-18 | (IN202441020271)Machine learning-powered system for unmasking fabricated faces | Makandar AzizJaveriya Syeda BibiKaman Shilpa |
IN202321014395 | (IN202321014395)2023-03-03 | (IN202321014395)Classification of deepfake images using a time distributed. | Kosqrkar Usha PrashantSakarkar Gopal |
IN202241005813 | (IN202241005813)2022-02-03 | (IN202241005813)Latent modalities for deception detection: a dl perspective | Salini YalamanchiliAlluri Bksp Kumar Raju |
IN202221028426 | (IN202221028426)2022-05-18 | (IN202221028426)Audiobook generation using deepfakes lip sync and translated text-tospeech system | Shinde SwatiShelar JayeshGhatole DipakPachpande MayankBhandari Dhanashree |
IV. GLOBAL REGULATIONS
USA
Amidst the recent news of Taylor Swift deepfake going viral, the bipartisan committee of US Senators have proposed the Disrupt Explicit Forged Images and Non-Consensual Edits (DEFIANCE) Act. The said proposed Act “would make sharing of nonconsensual, sexualized image generated with the help of Artificial Intelligence without the prior permission of the individual illegal”.[xiii]
Moreover, on March 21, 2024, Tennessee enacted the Ensuring Likeness, Voice, and Image Security (ELVIS) Act of 2024, to protect the rights of musicians and artists against unauthorized AI generated voices of artists and deepfakes.[xiv]
EU
The European Commission has introduced a new directive that addresses the issue of cyber-violence against women. The directive penalizes actions such as sharing intimate images without consent including deepfakes generated by implementing AI technology and online harassment. This directive is expected to make its debut in 2027.[xv]
UK
After passing the Online Safety Act in 2023 which made sharing of deepfakes illegal, the Ministry of Justice of England and Wales recently introduced a new legislation under which creation of “sexually explicit” deepfake images of individuals without their proper consent would be considered as a criminal offence. The new Law proposed in April, 2024 provides that any person who creates sexually explicit deepfake image is expected to face unlimited fine and criminal record. The proposed new legislation would be inserted as an amendment to the Criminal Justice Bill. [xvi]
INDIA
To suppress the dubious nature of deepfake technology, the Ministry of Electronic and Information Technology (MeitY) had issued a notice to major social media platforms and intermediaries, ensuring compliance with the existing IT rules the directive specifically addresses the growing concerns surrounding deepfakes and misinformation.[xvii]
MeitY Advisory
As per Rule 3(1) (b) provided under the IT Rules, intermediaries are required to make the users aware about their rules, regulations, privacy policy and user agreement in their preferred language. It further stated that it is the responsibility of the intermediaries to make sure that users are prevented from “hosting, displaying, uploading, modifying, publishing, transmitting, storing, updating or sharing any information related to the 11 listed user harms or content prohibited on digital intermediaries”[xviii]
The advisory also to ensure that platforms remove misinformation, false or misleading content and material impersonating others, including deepfakes.[xix]
The advisory also stated that, “The users must be informed about the provisions enshrined in the Indian Penal Code (IPC) 1860, the IT Act, 2000 and such other laws that may be attracted in case Rule 3(1) (b) is violated. Furthermore, the terms of service and user agreements must specifically emphasize as to how intermediaries/platforms are required to notify law enforcement authorities about legal violations falling under the relevant Indian laws applicable to the context.”[xx]
V. CONCLUSION
In the beginning deepfake videos were mostly being used to create humorous content[xxi] but over time the technology is being used by miscreants in a dark and grim manner for committing crimes against women, financial frauds, scams etc. While public figures like celebrities, politicians, world leaders have been the target audience for the misuse of the technology it cannot be denied that the negative use of the technology indiscriminately affects people from all walks of life making them susceptible to manipulation and exploitation.
As said, Life and Liberty has its own way to find space. Similarly, just as there is an antivirus for every virus bugged in a computer program, technology driven by human ingenuity can effectively combat the ongoing issue of deepfakes. The blend of Technology along with the slew of Regulatory frameworks and legislations leading to traceability, the rapid march of AI deepfakes can be arrested.
[i] The author is Managing Partner at S.S. Rana & Co.
[ii] The author is Senior Patent Associate at S.S. Rana & Co.
Swayamsiddha Das, Junior Associate at S.S. Rana & Co. has assisted in the research of this Article.
[i] Ian Sample, What are deepfakes – and how can you spot them?, The Guardian, (Apr.8, 2024 at 9:00 am), https://www.theguardian.com/technology/2020/jan/13/what-are-deepfakes-and-how-can-you-spot-them.
[ii] China may disrupt elections in India using AI, warns Microsoft, ETOnline, Economic Times, (Apr 8, 2024 at 9:30 am), https://economictimes.indiatimes.com/news/elections/lok-sabha/india/china-may-disrupt-elections-in-india-using-ai-warns-microsoft/articleshow/109135055.cms?from=mdr.
[iii] Vishnu Vijandran, The Intersection of Deepfakes and Intellectual Property Law: Applying the Concept of Passing Off, Aqran Vijandran, (Apr.8, 2024 at 9:45 am), https://www.aqranvijandran.com/blog/the-intersection-of-deepfakes-and-intellectual-property-law-applying-the-concept-of-passing-off.
[iv] Artist Rights Alliance, 200+ Artists Urge Tech Platforms: Stop Devaluing Music, Medium, (Apr. 9, 2024, at 1:30 pm) https://artistrightsnow.medium.com/200-artists-urge-tech-platforms-stop-devaluing-music-559fb109bbac.
[v] HT News Desk, Creator of Rashmika Mandanna Deepfake video arrested: Delhi Police, (Apr. 9, 2024 at 1:30pm), https://www.hindustantimes.com/india-news/main-accused-in-rashmika-mandanna-deepfake-case-arrested-delhi-police-101705740719655.html.
[vi] Anil Kapoor v. Simply Life India & Ors. CS(COMM) 652/2023.
[vii] Para 29. Anil Kapoor v. Simply Life India & Ors. CS(COMM) 652/2023.
[viii] Dr. Nivash Jeevanandam, Five interesting tools to detect deepfake in 2023, IndiaAI, (Apr.9, 2024 at 2:00pm), https://indiaai.gov.in/article/five-interesting-tools-to-detect-deepfake-in-2023.
[ix] Id at 6.
[x] Id at 6.
[xi] Id at 6.
[xii]Microsoft Launches Video Authenticator to detect deepfake media, Hindustan Times Tech, (Apr.10, 2024 at 9:30 am) https://tech.hindustantimes.com/tech/news/microsoft-launches-video-authenticator-to-detect-deepfake-media-71599114673370.html.
[xiii]Taylor Swift AI deep fakes: The US’ DEFIANCE act to curb AI deep fake porn, Times of India (Apr.10, 2024, at 1:00pm),http://timesofindia.indiatimes.com/articleshow/107529367.cms?utm_source=contentofinterest&utm_medium=text&utm_campaign=cppst.
[xiv] Kanishka Singh, Tennessee becomes first US state with law protecting musicians from AI, Reuters, (Apr.10 2024, at 2:30pm) https://www.reuters.com/legal/tennessee-becomes-first-us-state-with-law-protecting-musicians-ai-2024-03-21/.
[xv] European Union proposes law to criminalize deepfake, Times of India, (Apr. 11 2024 at 1:00pm), http://timesofindia.indiatimes.com/articleshow/107529367.cms?utm_source=contentofinterest&utm_medium=text&utm_campaign=cppst.
[xvi] Christy Cooney, Creating sexually explicit deepfakes to become a criminal offence, BBC News, (Apr.16, 2024 at 11:00 am), https://www.bbc.com/news/uk-68823042.
[xvii] Ministry of Electronics & IT, MeitY issues advisory to all intermediaries to comply with existing IT rules, (Apr.10, 2024 at 12:30 pm), https://pib.gov.in/PressReleaseIframePage.aspx?PRID=1990542#:~:text=On%2017th%20November%2C%20the%20Prime,and%20amended%20in%20April%202023.
[xviii] Id at 16.
[xix] Anuradha Gandhi & Isha Sharma, MeitY’s Advisory Unveiled to tackle Deepfake menace, SS Rana & Co., (Apr. 11, 2024 at 1:30pm), https://ssrana.in/articles/meitys-advisory-unveiled-to-tackle-deepfake-menace/.
[xx] Id at 14.
[xxi]Rohit KVN, What are deepfakes and how to spot them, Deccan Herald, (Apr.8, 2024 at 10:00 am), https://www.deccanherald.com/technology/what-are-deepfakes-heres-how-to-spot-them-2760185.
Comentarios