Deepfakes - A Comprehensive Guide | UPSC Notes

Decoding Deepfakes

The term deepfake refers to digital media – whether it be video, audio, or images – that have been edited and manipulated using Artificial Intelligence . This artificial media content is termed as ‘deepfakes’, and it represents a hyper-realistic form of fabrication.

• The term is a fusion of ‘deep learning’ and ‘fake’. Deep learning is a subset of machine learning that is based on Artificial Neural Networks (ANNs).
• Artificial Intelligence (AI) created synthetic media or deepfakes can have positive uses in fields like film production, education, artistic expression, and criminal forensics.
• However, deepfake technology is also being exploited by cybercriminals for phishing attacks, financial fraud, identity theft, misinformation, and even political disturbances.

The term ‘deepfakes’ was coined in 2017 by a Reddit user named ‘deepfakes’, who posted manipulated explicit videos of celebrities.

The Working of Deepfakes

A deepfake video can convincingly show a person performing an action or speaking about a subject. The creation of these videos involves superimposing multiple images of the target person onto an existing face from different angles.

• Deepfake software developers typically use Generative Adversarial Networks (GANs).
• A GAN comprises two neural networks – the generator, which creates the fake media, and the discriminator, which determines whether the media is real or fake.
• Whenever the discriminator accurately detects a fake, it provides valuable feedback to the generator, helping it improve its future deepfakes.
• The creation of deepfakes generally requires a large amount of visual data about the target. This makes high-profile figures like celebrities and politicians, whose images and videos are readily available online, easy targets. Consequently, deepfakes pose a significant threat to the reputations of such individuals.

The Dark Side of Deepfakes

• One of the most alarming uses of deepfake technology is in the creation of synthetic adult content, also known as deepfake pornography.
• With the power to create and publish highly realistic videos, there is immense potential for harm to individuals, organizations, and societies.
• In some cases, an uncomfortable truth may be dismissed as a deepfake.
• Non-state actors can exploit deepfakes to incite anti-state sentiments among the populace.
• Deepfakes pose a significant threat to democratic processes like elections, where they can be used to spread misinformation and propaganda.

Countering Deepfakes

Detection – Developing technologies to identify deepfakes

• Major tech companies like Google, Meta, and Microsoft have voiced their opposition to deepfake technology and are working on tools to detect them.
• Microsoft, for instance, is developing a new anti-deepfake technology called Microsoft Video Authenticator to combat misinformation.

Transparency

• To foster trust, social media platforms have introduced “verified accounts”.
• One proposed solution to counter deepfakes involves the use of Non-Fungible Tokens (NFTs).

Regulations

• In 2019, about a dozen Federal and State Level Bills were introduced in the US to regulate deepfakes. These laws cover a wide range of issues, from criminalizing the use of a woman’s likeness in a pornographic film without her consent (Virginia Law), to the appropriate use of a deceased person’s data (New York Law), and dealing with cheap fakes (low-tech digital frauds not requiring AI).
• A comprehensive regulatory framework could potentially deter the creation and dissemination of malicious deepfakes.

Education

• Media literacy for consumers is the most potent weapon against disinformation and deepfakes.
• The more the public is aware of the technology and its applications, the better they can critically evaluate the media they consume and exercise caution when necessary.

 

Related Links
Information Technology Act, 2000	Cybercrime
Internet of Things	Cybersecurity
Digital Personal Data Protection Bill, 2022	Science & Technology Notes For UPSC