In the ever-evolving realm of artificial intelligence, deepfakes emerge as the startling frontier, artfully blending faces and voices to craft digital mirages so convincing they challenge our very perception of reality. Imagine a world where digital avatars of ourselves can be manipulated to perform actions or utter words we’ve never said, blurring the lines between fiction and fact.
Deepfakes pose an increasing challenge due to their potential in disseminating false information and propaganda, as well as threatening or extorting individuals. They have found applications in numerous domains such as political campaigns, extortion, explicit content, governance, the arts, theatrical performances, marketing campaigns, entertainment sectors, internet humor, and various social media platforms.
Social Media and Government role in Deep Fakes:
Social media enthusiasts can employ AI-driven facial generation tools to mask their identity in photos shared by others. Some platforms, such as Facebook and Instagram, grant users the choice to accept or decline photo tags. It’s essential to also concentrate on: (a) Transforming fact-checking into a commercial service. (b) Establishing regulatory measures. (c) Developing organized fact-checking databases via Claim Review. (d) Regionally collaborate.
Social Engineering Crimes using Deep Fakes:
Deepfakes have the potential to play a role in social engineering offenses, where culprits employ cunning and deceit to coax individuals into revealing confidential data or undertaking specific tasks.
- A practical illustration of deepfakes in social engineering involves fabricating a video where a prominent executive or authority figure, like a CEO or a government leader, solicits confidential details or a money transfer. When such a video reaches an unsuspecting employee or another recipient, they might not discern its inauthenticity and might inadvertently fulfil the request.
- Another example is creating a fake video of a friend or family member asking for money or other financial help.
- Deepfakes can also be used in impersonation attacks, in which an attacker creates a fake video or audio of an individual, such as a celebrity or public figure, and it can mimic individuals to illegitimately access confidential data or valuable assets.
- It is important to be aware of these threats and take steps to protect yourself and your organisation from deepfake-based social engineering attacks. This entails maintaining a cautious approach towards unanticipated demands for data or funds, confirming the authenticity of the individual placing the request, and notifying any dubious occurrences to the relevant officials.
Fact Checking Deep Fakes:
Detecting deepfakes is a daunting task given the rapid advancements and increased sophistication of the underlying technology. Nonetheless, there are multiple measures one can adopt to ascertain the genuineness of a video:
- Check the source: Look for information about the video’s origin and try to verify that the source is reputable and credible.
- Check other sources: Look for other sources that report on the same video or story and compare the information to see if it matches up.
- Look for inconsistencies: Check for any inconsistencies or errors in the video, such as unnatural movements or expressions, mismatched audio, or inconsistencies in the lighting or shadows.
- Use specialised tools: There are tools available that can analyse videos and detect signs of manipulation, such as the “Deepfake Detection Model” which is a specialised software that can detect deepfakes.
- Check the context: Consider the context of the video and the circumstances under which it was created. Is it plausible that the events in the video could have occurred?
- Verify with the person: If the video features a specific person, try to reach out to that person and verify the authenticity of the video.
- Log on to https://factly.in/ or https://www.boomlive.in/ or https://www.altnews.in/ or any other official International Fact Checking Network Members to validate about Deep Fakes.
Spotting Deep Fakes:
- Look for unnatural movements or expressions: Deepfakes often struggle to perfectly replicate the subtle movements and expressions of real people, so look for unnatural or exaggerated movements in the video.
- Check the audio: Deepfakes may not perfectly match the audio with the video, so listen for any inconsistencies or errors in the audio.
- Check the lighting and shadows: Deepfakes may not accurately replicate the lighting and shadows in a scene, so look for inconsistencies or errors in these elements.
- Check the background: In a deepfake video, the background may be out of focus, or not match the movement of the subject.
- Check the metadata: Some deepfake videos may have metadata that indicates that the video is a deepfake or has been manipulated.
- Pay attention to blinking, check if the person blinking enough or too much.
- Ask the person to way the hand while talking.
- Ask the person to move the position of the face while talking.
- Use specialised software: There are software like “Deepfake detection model” which helps to detect deepfakes. https://deepware.ai, https://deepfakedetector.ai, https:// www.firstdraftnews.org and http://www.aifoundation.com
Conclusion:
Deepfake technology is constantly evolving, and it is becoming more difficult to spot deepfakes. Therefore, it is essential to remain sceptical and verify the authenticity of any video or image or a document before accepting it as true.
-Anil Rachamalla
