This guide is general education, not legal advice. In short, the US, UK and EU now all target intimate AI fakes of identifiable people, in different ways and on different timelines. For users of AI video apps, the safe rule is simple: never use a real person’s photo or likeness.
United States: the TAKE IT DOWN Act
Congress passed the TAKE IT DOWN Act (S. 146) on April 28, 2025. It was signed on May 19, 2025, per a Congressional Research Service report.
The law makes it a crime to knowingly publish intimate images of an identifiable person. It covers real images and “digital forgeries,” of adults and minors. Threats to publish are also covered.
Penalties
Per the CRS report, publishing real images carries up to 2 years in prison when the person is an adult. The maximum is 3 years when the person is a minor.
Threats involving digital forgeries carry up to 18 months for adults and 30 months for minors, per the same report. It sets out the full penalty structure.
Platform duties
Covered platforms must remove reported images “no later than 48 hours” after a valid request, per the CRS report. They had until May 19, 2026, to set up that process.
The Federal Trade Commission enforces these duties. The FTC began enforcing on May 19, 2026, and opened a complaint site at TakeItDown.ftc.gov.
On May 20, 2026, the FTC sent warning letters to 12 “nudify” companies. It noted penalties of up to $53,088 per violation.
United Kingdom: sharing, creating and requesting
Sharing (section 66B)
Section 188 of the Online Safety Act 2023 inserted section 66B into the Sexual Offences Act 2003. It covers sharing images that “show, or appear to show” a person in an intimate state.
The words “appear to show” mean it can reach images that were faked. Section 66B has been in force since January 31, 2024, per legislation.gov.uk.
Creating and requesting (sections 66E and 66F)
Section 138 of the Data (Use and Access) Act 2025 added two newer offenses. Section 66E covers creating “purported intimate images” of an adult without consent.
Section 66F covers requesting that such an image be created. That matters for app users, because asking for a fake of a real adult can itself be an offense.
Both are summary-only offenses, the less serious category heard in the lower courts. They have been in force since February 6, 2026, under SI 2026/31.
We have not confirmed which parts of the UK these sections cover. If you live in Scotland or Northern Ireland, check local law.
European Union: the AI Act and the 2024 directive
AI Act transparency (Article 50(4))
Under Article 50(4) of the AI Act, deployers of deepfakes must disclose that the content was generated by AI. This duty applies from August 2, 2026.
The Digital Omnibus bans
The Digital Omnibus on AI is Regulation (EU) 2026/1744. It was adopted on July 8, 2026, and published in the Official Journal on July 24, 2026.
It entered into force on July 27, 2026, per the European Commission.
It adds new bans in Article 5(1)(ba) and (bb) of the AI Act. These cover AI systems that generate non-consensual intimate content of identifiable people, and systems that generate child sexual abuse material.
The bans apply from December 2, 2026. The European Parliament described the deal as including a ban on nudifier apps.
Generative systems on the market before August 2, 2026, have until December 2, 2026, to meet the Article 50(2) marking duty. Some sources still give February 2, 2027, but that date came from the proposal.
Directive (EU) 2024/1385
Directive (EU) 2024/1385 was adopted on May 14, 2024. Article 5(1)(b) requires member states to make certain fakes a crime.
It covers producing or manipulating material that makes it appear a person is engaged in sexually explicit activity. Making it accessible is covered too, where done without consent and likely to cause serious harm.
Article 10(4) requires a maximum penalty of at least one year. Article 23 requires prompt removal of such material.
Member states must transpose the directive by June 14, 2027. Until then, national laws vary.
Key dates at a glance
- January 31, 2024: UK section 66B sharing offense in force.
- May 14, 2024: EU Directive 2024/1385 adopted.
- May 19, 2025: US TAKE IT DOWN Act signed.
- February 6, 2026: UK sections 66E and 66F in force.
- May 19, 2026: FTC begins enforcing the platform removal duty.
- August 2, 2026: EU AI Act Article 50(4) deepfake disclosure applies.
- December 2, 2026: EU bans on non-consensual intimate content systems apply.
- June 14, 2027: EU deadline to transpose Directive 2024/1385.
How Google Search treats explicit fakes
On July 31, 2024, Google’s Emma Higham announced changes to how Search handles explicit fakes. The Search Central changelog logged the update the same day.
- After a successful removal, Google aims to filter explicit results on similar searches about that person.
- It also removes duplicates of images it has removed.
- Ranking updates lower explicit fakes on searches for a person’s name. Google said exposure to such results was “reduced… by over 70%.”
- Sites with a high volume of removals are demoted in rankings.
How to request removal
Google’s removal help page sets out who qualifies. You must be identifiable, the content must be fake and nude or sexual, and it must have been shared without consent.
You, or someone representing you, file the form with the URLs and screenshots. Removal from Search does not delete the content from the site that hosts it.
What this means for users of AI video apps
- Stick to fictional adults. Consent from a real person is hard to prove, and the laws above focus on identifiable people.
- Never upload someone else’s photo. That includes using it as a “reference” or a starting frame.
- Do not ask others to make fakes. In the UK, requesting one can itself be an offense.
- Do not threaten to share. The TAKE IT DOWN Act covers threats to publish as well as publishing.
- Label AI content. In the EU, deployers of deepfakes must disclose AI generation from August 2, 2026.
The apps we profile set their own rules too. Promptchan requires “explicit and verifiable consent” for real people. OurDream AI says it does not host real-person likeness.
Candy AI has zero tolerance for deepfakes and does not allow uploads. Secrets AI bans intimate imagery of real individuals.
Kupid AI‘s terms do not address real-person uploads. Its anti-trafficking policy says it reviews content that may involve a real person.
Our consent policy sets the line for this site. The Prompt Lab accepts only text descriptions of fictional adults and refuses real people and minors.
If you run models locally, no app enforces these rules for you. Our Wan 2.2 setup guide covers that in its safety section.
Check current plans on OurDream AI
Explicitly bans real-person likeness and lookalikes in its prohibited content policy.
Where to get help
If someone has made or shared an intimate fake of you, these options may help.
- US federal complaint: the FTC’s complaint site for the TAKE IT DOWN Act is TakeItDown.ftc.gov.
- Google Search removal: use the form on Google’s removal help page.
- StopNCII.org: a widely known service that helps adults stop intimate images from being shared on partner platforms.
- NCMEC’s Take It Down service: for images made when the person was under 18. It is separate from the FTC complaint site.
- Police and a lawyer: especially for threats, extortion or anything involving a minor.
Sources and dates
Facts checked against the linked official pages on October 31, 2026. Laws and deadlines in this area change quickly, so check the linked sources for updates.