Deepfake Technology

Deepfake Technology

 

According to one firm, the number of fake videos on the internet increased from 14,678 in 2019 to 145,277 in June of the following year. The use of technology for campaigning by Indian politician Manoj Tiwari raised questions.

The makers of the iconic American comedy series "South Park" released the viral video "Sassy Justice" last year, which featured deep-faked versions of Trump and Mark Zuckerberg. In an interview, they said that while worry over deepfakes may have taken a back seat to pandemic-related concerns, the matter is still worth debating.

  

What is "deepfake technology"?

 

It is a technique for modifying movies, photos, and audio that makes use of powerful computers and deep learning. The name is a mix of two terms: "deep learning" and "false," which makes use of powerful computers and deep learning. The name is a mix of two terms: "deep learning" and "false." Deep learning is a subset of machine learning that is based on artificial neural networks (ANNs).
Artificial intelligence (AI)-created synthetic media, often known as "deep fakes," provide a variety of advantages in sectors such as film production, education, creative expression, and criminal forensics.
Deepfake technology is being used by cybercriminals to carry out phishing attacks, financial fraud, identity theft, information manipulation, and political instability.

 

The term "Deepfake" was first used in 2017 by an unidentified Reddit user with the username "Deepfakes." This user took advantage of DNN (deep neural networks) based on a face-swapping algorithm to generate and share pornographic films.

 

AI algorithms transfer a person's words, head movements, and expressions. The technology for creating deepfakes is available to ordinary users; several software tools, such as FakeApp, DFaker, Faceswap-GAN, and Deepface Lab, are freely available on GitHub.

 

What is "Deep Synthesis"?

 

• Deep synthesis is described as the use of technology such as deep learning and augmented reality to construct virtual scenarios using text, photos, audio, and video.
• Deepfakes are one of the most notorious implementations of the technology, in which synthetic media is used to replace one person's face or voice with another's.
• Deepfakes are becoming more difficult to detect as technology advances. It is used to generate false news, manufacture celebrity porn videos, perpetrate financial fraud, and so forth.

 

 

How does deepfake technology function?

The deep fake video seems to be original footage, with the person doing an action or speaking about a topic. Moreover, many photos of the targeted individual from various perspectives are placed on the real face while generating such false recordings.

• The process of creating a deep-fake video involves using a Generative Adversarial Network (GAN) to combine and superimpose existing images and videos onto source images or videos.
• GAN is made up of two neural networks: the generator, which generates false media, and the discriminator, which evaluates if the media is genuine or fake.
• Every time the discriminator properly identifies the material as fake, the generator gains crucial information about how to improve future deep fakes.
• Deepfake-generation techniques often need a large number of pictures and videos of the targets.
• Because of the vast data sets available on the internet, high-profile individuals such as celebrities and political leaders become easy targets. As a result, thorough forgeries can gravely ruin the reputation of any high-profile individual.
  
  

 

 

What are "deep-fake" threats?

 

• 96% of deepfakes are pornographic videos that humiliate women and cause emotional harm.
• Misuse by non-state actors (terrorist organisations) to create anti-state sentiment among the general public.
• Deepfak may portray a person as engaging in deviant behaviour, causing social conflict and increasing division.
• Election-related democratic procedures can be undermined with the use of deep fakes.
• Deepfakes have become a tool of disinformation, destabilising the political process, pitting citizens against each other, and allowing politicians to manipulate the truth.
• Fakes are being weaponized, with fake news and alternative-facts narratives being used to reject legitimate media and the truth.

 

What steps are nations taking to combat deep fakes?

 

CHINA:

• China will implement a first-of-its-kind law on "deepfakes" in January 2023, tightening control over online material.
• CAC (China's Cyberspace Administration) has published laws prohibiting the use of "deepfakes" without the subjects' permission, as well as depicting or uttering anything that might be regarded as contrary to national interest.
• The laws, formally titled "The Administrative Provisions on Deep Synthesis for Internet Information Services," were created in response to government fears that breakthroughs in AI technology may be exploited by bad actors to operate frauds or defame individuals by misrepresenting their identities.
  
  

 

EUROPEAN UNION

 

The European Union has issued an amended "Code of Practice" requiring digital companies like Google, Meta, and Twitter to take steps to combat deep fakes on their platforms.

 

CANADA

 

While Canada lacks rules to combat deep fakes, it is conducting some of the most cutting-edge AI research with a variety of local and international entities.
In addition, Canada is a member and leader in a number of similar global projects (Paris Call for Trust and Security in Cyberspace, Global Partnership on AI), which may be used to develop deep-fake policies in a variety of sectors.

 

United States of America

 

USA introduced The bipartisan Deepfake Task Force Act would help the Department of Homeland Security (DHS) combat deepfake technology by conducting an annual deepfake study.

 

INDIA

 

There are no legal restrictions in India regarding the use of deep fake technologies. However, certain laws can be addressed for abusing technology, such as copyright infringement, defamation, and cybercrime.

 

• In the case of Justice KS Puttaswamy v. Union of India The right to privacy has been recognised as a fundamental right in India.
• Currently, there are relatively few laws under the Indian Penal Code (IPC) and the Information Technology Act of 2000 that may be used to deal with the malicious usage of deep fakes. Defamation is a crime under Section 500 of the Indian Penal Code. 
• Sexually explicit material is prohibited under Sections 67 and 67A of the Information Technology Act.

 

• During an election period, the Representation of the People Act, 1951,India has no legal restrictions circulation of false or misleading material regarding candidates or political parties.

 

• According to regulations established by the Indian Election Commission, all political ads on electronic media, including TV and social media websites, must first have preapproval from registered political parties and candidates. This is done to help assure the accuracy and impartiality of the commercials.

 

How to deal with "deepfake" technology

 

• Building techniques for detecting deep fakes : For example, Google, Meta, and Microsoft have all criticised deep fake technology and are developing methods to detect it. In addition, Microsoft is developing a new anti-deep fake technology called Microsoft Video Authenticator to combat disinformation.
• Create transparency: Social media corporations introduced "verified accounts" to instil confidence.
  One proposal for combating deep fakes suggests using NFT (non-fungible tokens) as a viable solution.

• Literacy and Awareness: Consumer media literacy is the most effective strategy for combating misinformation and deep fakes. The greater the public's understanding of technology and its applications, the more they will be able to think critically about the media they consume and exercise care when necessary.
• Regulation: Meaningful policies were developed in collaboration with the technology industry, civil society, and government to disincentivize the creation and dissemination of harmful deepfakes.
• Easy and accessible technology : solutions for detecting deep fakes, authenticating information, and amplifying authoritative sources that are simple to use and accessible

 

 

Way Forward

 

• Separate legislation should be enacted by the Union government to address both the criminal use of deepfakes and the broader issue of AI.
• The legal framework for AI in India is insufficient to meet the numerous concerns that have evolved as a result of AI algorithms.
• Deepfakes may allow someone to reject the legitimacy of actual information, especially if it depicts them engaged in improper or illegal activity, by saying that it is a deep fake.

• We must add rigor and accountability to AI technology development, implement ethical rules, perform frequent audits for fairness, uncover and fix biases, and safeguard privacy and security. 

• This issue may potentially be addressed by the planned Digital India Bill.