AI Deepfakes - What is actually real?
AI deepfakes. What are they, why are they suddenly everywhere, and how can you tell what’s real?
Today’s post is all about deepfakes—the synthetic media that’s blurring the line between real and fake. We’ll break down what deepfakes are, the risks they pose, and how you can protect yourself and your community.
Generated using Gemini 2.5 Flash
What Are Deepfakes?
Deepfakes are synthetic audio, images, or videos created using AI, specifically deep learning. These tools can mimic a person’s face, voice, or expressions so convincingly that it becomes difficult to tell what’s real and what’s not. As Dr. Manjeet Rege, Director of the Center for Applied Artificial Intelligence, explains, “They’re made by training neural networks on large datasets of images, videos, and audio recordings to produce convincing forgeries that can spread misinformation if used maliciously.
The Risks of Deepfakes
Let’s talk about the dangers. Deepfakes can:
Spread Misinformation: Fake videos of politicians or public figures saying things they never said can influence elections, incite violence, or damage reputations.
Invasive Privacy: Deepfake pornography—often targeting celebrities or private individuals without consent—has become a significant issue. In fact, 96% of deepfake videos online are pornographic, and the numbers are rising.
Enable Scams and Fraud: Imagine a CEO’s voice being faked to authorize a fraudulent transaction, or a scammer using your likeness to trick your loved ones.
Undermine Trust: As deepfakes become more convincing, people start doubting everything they see and hear online, which erodes trust in media, institutions, and even personal relationships.
Legal and Regulatory Risks: Sharing or creating deepfake content, especially explicit or defamatory material, can lead to lawsuits or criminal charges in many regions, as new laws are introduced to combat these threats.
Real-World Examples
Recent incidents highlight the risks:
Fake nude images of celebrities like Taylor Swift have gone viral, sparking public outrage and legislative action.
Business leaders and politicians have been targeted with deepfake videos and audio, sometimes resulting in financial losses or reputational crises.
Children and teens are increasingly at risk, with cases of classmates creating inappropriate deepfakes for bullying or revenge.
Realistic Advice to Protect Yourself from Deepfakes
1. Limit What You Share Online
Be selective about the photos, videos, and personal information you post, especially high-resolution images or videos that deepfake creators can easily manipulate.
Adjust your privacy settings on all social media and cloud platforms to restrict access to trusted friends and contacts only.
2. Use Watermarks and Image Protection
Add digital watermarks or subtle filters to your photos and videos before sharing them online. This makes them harder to use for deepfake creation and easier to trace if misused.
Consider using artistic filters or cropping images to obscure facial features, which will make it more difficult for AI to extract your likeness.
3. Strengthen Your Digital Security
Use strong, unique passwords for each of your accounts and enable multi-factor authentication (MFA) wherever possible.
You can keep your devices and software updated to fix any security vulnerabilities that could be exploited to access your media.
Install reputable antivirus and anti-malware software to protect against hacking attempts and malware that could steal your images or videos.
4. Monitor Your Digital Presence
Set up Google Alerts or use identity monitoring services to get notified if your name or images appear in unexpected places online.
Regularly search for your name and likeness to catch unauthorized use early.
5. Learn to Spot Deepfakes
Watch for typical signs such as unnatural blinking, odd facial movements, mismatched audio, inconsistent lighting, or blurry/distorted areas in videos and images.
Be skeptical of sensational or out-of-character content, especially if it involves public figures or comes from unverified sources.
Use AI-powered detection tools like Deepware.ai or Sightengine.com to analyze suspicious content.
6. Respond Quickly if Targeted
If you discover a deepfake of yourself, document everything: take screenshots, save links, and record any related communications.
Report the content to the platform hosting it and, if necessary, to law enforcement or relevant authorities.
Consult legal experts specializing in data privacy or cyber law for advice on takedown procedures and potential legal action.
Seek emotional support if needed; being the target of a deepfake can be distressing, and support groups or mental health professionals can help.
7. Stay Informed and Educate Others
Keep up with the latest trends in deepfake technology and scams, and share this knowledge with friends, family, and colleagues to build a more vigilant community.
Deepfakes are a powerful technology with both creative and destructive potential. The best defense is to stay informed, vigilant, and proactive. For more information, check out the sources below.
Thanks for listening to Tech Break by Friday. Subscribe for more insights on how tech shapes our world. See you next time!
Sources:
[1] https://ppl-ai-file-upload.s3.amazonaws.com/web/direct-files/collection_ae65a61c-b8e5-4986-8192-dff3d2341a5d/92f759f5-2193-48ca-b8d0-277c511ace02/paste-2.txt
[2] https://www.staysafeonline.org/articles/how-to-protect-yourself-against-deepfakes
[3] https://enfortra.com/10-tips-to-protect-yourself-from-deepfakes/
[4] https://www.scoredetect.com/blog/posts/legal-remedies-for-deepfake-victims-guide
[5] https://www.kaspersky.com/resource-center/threats/protect-yourself-from-deep-fake
[6] https://www.charityexcellence.co.uk/ai-deepfake-video-and-images/
[7] https://www.mcafee.com/ai/news/deepfake-defense-your-8-step-shield-against-digital-deceit/
[8] https://www.ndtv.com/offbeat/how-to-protect-your-images-and-videos-from-being-targeted-by-deepfake-tools-5288987
[9] https://www.globalsecuritymag.com/6-tips-to-help-you-protect-yourself-from-risks-stemming-from-deepfakes.html
[10] https://kpmg.com/us/en/articles/2025/protect-against-deepfakes.html
[11] https://connectsafely.org/ask-trish-help-im-the-victim-of-non-consensual-sexual-deepfake/
[12] https://english.elpais.com/technology/2024-02-28/what-are-deepfakes-their-risks-and-how-to-spot-them.html
[13] https://www.youtube.com/watch?v=AvxJJSPI99E
[15] https://www.iheart.com/podcast/269-voices-for-voices-89825951/episode/the-scary-future-of-deepfake-technology-202696020/
[16] https://rss.com/podcasts/cit-techforbusiness/643774/
[17] https://rtp.fedsoc.org/podcast/deep-dive-ep-168/