With improvements in technology related to artificial intelligence (AI), deepfakes are becoming common on the internet which include pictures, audio or videos.

A combination of AI and machine learning used to manipulate or generate visual and audio content with a high potential to deceive.
It is part of deep synthesis, utilizing AI to create realistic text, images, audio, and video.

Applications and Misuse:

Deepfakes are used for generating fake news, committing financial fraud, and more.
They manipulate existing content with AI to create convincing forgeries.
Machine learning algorithms make deepfakes more sophisticated than older photo-editing techniques.
Used to create false portrayals of political figures and fabricate events.

Positive Uses of Deep Learning:

Ethical and Social Implications:

India’s Stance on Deepfakes:

India currently lacks laws specifically addressing deepfakes.
Sections of the Information Technology (IT) Act and Indian Penal Code (IPC) that deal with defamation and explicit content may apply.
Further, the Personal Data Protection Bill may offer indirect protections.

International Responses:

European Union: Updated the Code of Practice on Disinformation to combat deepfake spread.
United States: The Deepfake Task Force Act helps Department of Homeland Security (DHS) to fight deepfake technology.
China: Comprehensive regulations demand labelling, traceability, consent, and adherence to moral standards.

Strategies for Combating Deepfakes:

AI Fact-Checking: Social Media should utilize AI for identifying and flagging deepfakes.
Blockchain Verification: Implement blockchain to record digital media’s creation and edits for authenticity.
Deepfake Mitigation Fund: Financial support for those affected by deepfakes.
Legislative Actions: Introduction of Deepfake Accountability Acts to enforce responsible creation and distribution.
Punitive Measures & Education: Laws to punish misuse and campaigns to raise public awareness on deepfakes.

Identifying ‘Deepfake’ Videos

· Deepfakes may show irregular eye movements or gaze patterns that do not align with the overall body language or spoken words.

· Look for inconsistencies in colour tones and lighting on the subject’s face compared to the environment which could indicate manipulation.

· Compare the audio with visual elements; AI-generated audio might contain subtle flaws that reveal it is not authentic.

· Distorted or unnatural body movements, or incorrect limb proportions, can be a sign of a deepfake.

· Watch for facial expressions that are exaggerated, poorly synced with speech, or otherwise out of context.

· Distortions or misplacements of facial features might signal the video has been altered.

· Unusual body positions or movements that seem implausible or awkward could be evidence of a deepfake video.

· Use a screenshot of the video and run a reverse image search (e.g., via Google Images) to find the source or verify the original content.