Full Report
The UK government is cracking down on the generation of sexually explicit deepfakes in a bid to protect women and girls
Analysis Summary
# Regulation/Compliance: UK Ban on Creation of Explicit Deepfakes
## Overview
This developing regulation in the UK aims to criminalize the *creation* of sexually explicit deepfakes, adding to existing offenses related to sharing or threatening to share such content under the Online Safety Act 2023. The prohibition also covers the act of taking or recording an intimate photograph or film without consent, regardless of whether the intent is sexual gratification or causing alarm, distress, or humiliation.
## Key Details
- **Issuing Authority:** UK Government (Legislation being enacted/amended).
- **Effective Date:** Not explicitly stated for the creation ban, but the sharing/threat aspects are covered by the Online Safety Act 2023.
- **Jurisdiction:** United Kingdom.
- **Status:** Announced plans/Developing Legislation (The creation ban is being added to existing offenses under the Sexual Offences Act 2003, via amendments related to the Online Safety Act 2023).
## Requirements
### Mandatory Requirements
1. **Criminal Offense:** It will be a criminal offense to *create* sexually explicit deepfakes.
2. **Extended Scope of Offense:** The law will also criminalize the taking or recording of an intimate photograph or film without consent, irrespective of the intent (i.e., whether for sexual gratification or to cause alarm, distress, or humiliation).
3. **Existing Obligation:** Organizations remain subject to existing provisions under the Online Safety Act 2023 regarding the sharing or threatening to share intimate images, including deepfakes.
### Recommended Practices
1. **Review Content Policies:** Organizations dealing with user-generated content should proactively review and update policies to specifically address the creation, sharing, and hosting of explicit deepfakes based on the forthcoming specific legislation.
2. **User Education:** Educate users about the new criminal liability associated with the creation and distribution of non-consensual intimate imagery (including deepfakes).
## Affected Organizations
- **Industries:** Primarily social media platforms, content hosting services, file-sharing platforms, and any entity facilitating the creation or distribution of digital media (given the context of the Online Safety Act provisions).
- **Organization Size:** Applicable regardless of size, as it establishes baseline criminal law.
- **Geographic Scope:** United Kingdom.
## Compliance Timeline
- **N/A (Existing):** Sharing or threatening to share intimate deepfakes criminalized by amendments in the **Online Safety Act 2023**.
- **TBD (Future):** Specific date for when the **creation of explicit deepfakes** becomes a specific criminal offense.
- **Final deadline:** Full compliance with the new creation ban requires adherence once the relevant legal provisions come into force.
## Implementation Guidance
### Assessment Phase
- **Policy Gap Analysis:** Assess current terms of service and content moderation policies against the specific prohibitions concerning the *creation* of explicit deepfakes and non-consensual intimate image recording.
### Implementation Phase
- **Moderation Updates:** Deploy or refine technological and human moderation processes to detect and stop the creation or uploading of materials identified as explicit deepfakes, aligning with the broader obligations under the Online Safety Act regime.
### Validation Phase
- **Testing Protocols:** Establish testing protocols to ensure content pipelines prevent the creation/upload of prohibited material, consistent with general UK digital safety framework expectations.
## Technical Requirements
*The article focuses on legal classification rather than specific technical mandates for platforms, but achieving compliance necessitates:*
1. **Content Detection:** Implementing or upgrading AI/ML tools capable of identifying synthetic/manipulated explicit imagery (deepfakes).
2. **Data Handling Protocols:** Strict controls over user-submitted content to prevent unauthorized recording or creation involving intimate material.
## Penalties & Enforcement
- **Fines:** Not explicitly detailed for the creation offense in the excerpt.
- **Other Consequences:** Perpetrators found guilty of creating explicit deepfakes face **up to two years behind bars** (imprisonment).
- **Enforcement:** Will be enforced by UK authorities under criminal law, leveraging powers potentially established or augmented by the Online Safety Act framework.
## Related Standards
- **Online Safety Act 2023 (UK):** Provides the legislative framework under which these recent deepfake provisions are being introduced or strengthened.
- **Sexual Offences Act 2003 (UK):** The legislation being amended to incorporate these new criminal offenses related to intimate imagery.
## Resources
- **Official Documentation:** Reference the text of the **Online Safety Act 2023** and subsequent amendments related to the **Sexual Offences Act 2003**.
- **Guidance Documents:** Seek official guidance materials published by the UK government (e.g., DCMS or Ofcom) pertaining to the implementation of the Online Safety Act and Non-Consensual Intimate Image prohibitions.
- **Tools:** Deepfake detection tools and robust digital forensics capabilities.
## Practical Recommendations
1. **Immediate Action on Sharing:** Organizations must rigorously enforce existing policies against sharing or threatening to share intimate deepfakes, as this is already illegal under the OSA 2023.
2. **Prepare for Creation Liability:** Begin internal planning for how to manage liabilities or technical safeguards necessary to prevent the known *creation* of illegal synthetic media once the new legislation is finalized and enforced.
3. **Focus on Consent:** Ensure all processes related to capturing or processing intimate user imagery rigorously adhere to explicit, affirmative consent requirements, as the law broadened the scope of non-consensual recording offenses.