The first phase of voting for the Lok Sabha elections will be held on Friday (April 19). Over the past few weeks, there has been a flood of disinformation and manipulated media online.
two This week, a video of actor Aamir Khan went viral. Both were edited versions of Khan's promos for his popular TV show Satyamev Jayate. In one photo, Khan appears to be clearly endorsing the Congress party, while in the other he is seen talking about Nyay (justice), a key issue for the Congress in recent years, and the title of its manifesto (Nyay Patra or “Documents''). [for] justice').
Recently, actor Ranveer Singh also became a victim of deepfake technology after a doctored video criticizing Prime Minister Narendra Modi on issues of unemployment and inflation was shared widely. However, in the original clip, Ranveer was actually praising the PM.
Here's how these deepfake videos are created and how to spot them.
voice exchange technology
The AI detection tool isaar.ai, developed in conjunction with IIT Jodhpur, indicates that these videos were generated using 'voice exchange' technology.
As the name suggests, this refers to the process of using AI algorithms to alter or imitate an individual's voice. The technology also allows creators to change audio characteristics such as accent, tone, pitch, and speech patterns to make their videos more realistic.
There are currently several easy-to-use AI voice exchange tools available for free. Authors simply upload or record the audio sample they want to replace, then customize the settings to make the uploaded sample sound as realistic as possible.
How to spot a deepfake
Spotting a well-crafted deepfake isn't easy, but here are some tips to keep in mind as you scroll through social media, especially during election periods.
Check the source. Be wary of audio or video content from unfamiliar sources, especially if you believe it to be controversial or sensational. Verify the authenticity of questionable posts by cross-referencing them with trusted sources and trusted news organizations.
Listen for anomalies. Deepfake audio may exhibit subtle anomalies, such as an unnatural tenor in the voice, a slightly robotic speaking style, or irregular pauses. Listen carefully for these telltale signs of manipulated or synthetic audio.
Scrutinize visual content: Deepfake audio is often accompanied by manipulated visual content, such as altered video footage. Check both audio and visual elements for inconsistencies and inconsistencies. For example, if your lips don't move in sync with your speech, the video you're watching may be manipulated.
Stay informed: To be aware of the risks associated with deepfakes, it is important to stay informed about daily news and events. It is more difficult to fool people who are generally aware of what is happening around them.
Use AI voice detector. Some AI detectors are free to use, such as Optic's “AI or Not.” Uploading suspicious audio or video to a detector like this will tell you the authenticity of the content.
