Combat Against AI Deepfakes

[KitaYama]

Denmark is set to become the first European country to give people copyright control over their own image, voice, and facial features to combat AI deepfakes.

A new amendment to the copyright law, expected to take effect in fall 2025, will let citizens demand the removal of and seek compensation for unauthorized AI-generated portrayals, with exceptions for satire and parody.

Platforms that violate the law may face heavy fines, and Denmark plans to advocate for similar protections across the EU during its upcoming presidency.

Denmark to tackle deepfakes by giving people copyright to their own features

Amendment to law will strengthen protection against digital imitations of people’s identities, government says

www.theguardian.com

Danes Could Get Copyright to Their Own Image Under AI Bill

Denmark is considering giving people copyright control over their own likeness, voice, and features to combat deepfakes.

time.com

 

[The_Doc_Man]

I applaud their efforts. Not to be a nay-sayer, but in the USA our laws are SOMEWHAT different - not sure how much. The USA-based service providers (platforms) have some protections that would immunize them from responsibility for a member's action.

 

[KitaYama]

I applaud their efforts. Not to be a nay-sayer, but in the USA our laws are SOMEWHAT different - not sure how much. The USA-based service providers (platforms) have some protections that would immunize them from responsibility for a member's action.

I've been following this news for a while, and was very interested to see how they can achieve this. There are thousands of AI app on Google play that you can download and use free of charge to create these deepfake videos and stream them into social media. How can they track who did this?

 

[The_Doc_Man]

How can they track who did this?

It will depend on whether the app leaves behind a "fingerprint" of the system on which it was built, but if something is transmitted via the network, it is POSSIBLE (not at all certain) to find some traceback info on the sender of the fake message.

 

[KitaYama]

The USA-based service providers (platforms) have some protections that would immunize them from responsibility

I really like how European countries are fighting back against this hard time of fake images.

France passed a law requiring that any commercial image (e.g. ads, magazines) depicting models whose bodies have been digitally altered must carry a clear label reading “photographie retouchée” (“retouched photograph”). Non-compliance could result in fines up to €37,500 or even six months in jail.

Norway amended its Marketing Control Act to require that any paid advertising or influencer posts using retouched or filtered images of bodies be labeled. The prescribed label, often translated as “retouched person advertisement”, must be clearly visible.
Here's some from Norway :

 

[Isaac]

It sounds good at first blush, but if I'm being totally honest with you, I think Denmark can afford to do this being a tiny country and perhaps without as much at stake.

If the whole world did this, it would severely limit and cramp the use of AI. At least I feel that way, I'm not certain

 

[AccessBlaster]

Do we need AI to expose other AI? And if so, isn’t that ultimately a conflict of interest?

 

[KitaYama]

If the whole world did this

Almost all countries have started. For example

US

• Legislation:
  • Several states (like California, Texas, and New York) have laws targeting deepfakes in elections and pornography.
    California: Illegal to publish misleading deepfakes within 60 days of an election.
    Texas: Criminalizes the creation of deepfakes with intent to harm or defraud.
  • Federal proposals like the DEEPFAKES Accountability Act seek to mandate disclosure labels.
• Defense & Research:
  • The Pentagon and DARPA are funding tools to detect manipulated media.

---

China:

• Strict Regulation:
  • Since 2023, China requires all AI-generated content (including deepfakes) to be clearly labeled.
  • Platforms and creators must verify identities and report harmful content.
  • Deepfake tools are banned unless they serve a legitimate purpose.
• Enforcement:
  • Violators face heavy fines and criminal charges under internet and cybersecurity laws.