The proposal is to add a new right under Dutch law, similar to rights for performing artists. This right would let individuals control how their image or voice is used in deepfakes.

The law would allow people to stop others from making, sharing, or profiting from deepfakes of them without permission.

Question 1 of 2 for ChatGPT-4o: Please read the document "Deepfakes under a neighbouring right" and tell me what it says in easy to understand language

ChatGPT-4o’s answer:
The document "Deepfakes under a neighbouring right" suggests a new legal framework to give people control over unauthorized deepfakes of their image or voice. Here’s a summary of its key points in simpler terms:

1. Need for New Protections: Current laws in the Netherlands, like portrait rights and defamation laws, offer some protection against deepfakes, especially for living people. But these protections are fragmented and often don’t apply well to deepfakes involving deceased people.

2. Proposal for a Neighbouring Right: The proposal is to add a new right under Dutch law, similar to rights for performing artists. This right would let individuals control how their image or voice is used in deepfakes. The law would allow people to stop others from making, sharing, or profiting from deepfakes of them without permission.

3. Definition of Deepfake: A deepfake is defined as AI-generated content that realistically mimics a real person’s appearance or voice, and could be mistaken for genuine by an average viewer. This would exclude obvious parodies or cartoonish imitations.

4. Rights After Death: The proposed right would extend for 70 years after a person’s death, letting their family or a designated person continue to control how deepfakes of them are used.

5. Limitations and Exceptions: Certain uses of deepfakes, like satire or obvious fiction, could still be allowed if they meet transparency requirements (e.g., indicating that the content is AI-generated). This aligns with transparency rules in the European AI Act.

6. Criminal Penalties: Making or sharing deepfakes without permission would be punishable under this proposed law, adding a layer of criminal enforcement.

The proposal aims to establish stronger, clearer protections against deepfakes, especially focusing on maintaining personal dignity and privacy even after death.


Question 2 of 2 for ChatGPT-4o: Do you agree with the proposal and the arguments made? Is anything missing from the document?

ChatGPT-4o’s answer:
The proposal for a neighbouring right specific to deepfakes is a promising step toward addressing the complexities and harms of this technology. It aligns with existing protections for performing artists and provides a structure that emphasizes personal control over one's image and voice. Here’s an analysis, including some potential strengths, limitations, and missing elements:

Strengths and Merits of the Proposal

1. Clarity in Protection: By defining deepfakes within a legal framework, the proposal offers individuals (and their estates after death) explicit rights, making it easier for people to enforce control over unauthorized deepfakes.

2. Extended Postmortem Rights: Extending protections for 70 years posthumously acknowledges that a person's image and voice can hold value long after death. This is particularly relevant for public figures and artists who might otherwise have no legal recourse against misuse after their death.

3. Balance with Freedom of Expression: The document thoughtfully includes exceptions for satire, parody, and obvious fiction, provided that transparency requirements are met. This respects artistic and expressive freedoms, a critical balance for legal frameworks dealing with creative technologies.

4. International Potential: Framing this right as a neighbouring right aligns with EU standards and could influence broader EU-wide or international regulations, which is essential for dealing with deepfakes given their borderless nature.

Potential Gaps and Areas for Further Development

1. Specificity on Enforcement and Technology: The proposal could delve deeper into enforcement mechanisms, especially given the difficulty of tracking deepfakes across digital platforms. While transparency obligations help, they might be insufficient on platforms where deepfake content spreads quickly and without clear attribution.

2. Intermediary Liability: As intermediaries (e.g., social media platforms) play a significant role in the spread of deepfakes, the proposal could address their responsibilities more directly. For instance, should platforms be required to proactively detect or remove unauthorized deepfakes? Clear rules around intermediary liability would bolster enforcement.

3. Consideration of Consent Revocation: The proposal provides for licensing deepfakes with limitations, but it could expand on circumstances under which someone might revoke permission. For example, if an individual’s context changes, or if their likeness in a deepfake is associated with reputational damage, they may wish to withdraw previously given consent.

4. Guidelines on Permissible Uses: While it suggests that satire and parody are allowed if transparent, the proposal could provide guidelines for assessing whether deepfakes constitute fair use. This would give courts a clearer framework to address potential disputes and ensure deepfake rights do not overreach or unduly hinder freedom of expression.

5. Global Enforcement Mechanisms: Although intended for the Netherlands, a more detailed discussion on how to address cross-border misuse could strengthen its application. Deepfakes, especially unauthorized ones, are often produced and distributed internationally, so addressing international cooperation or potential harmonization with EU and global laws might make the right more enforceable.

Final Assessment

Overall, the proposal provides a strong starting point for tackling the unique challenges posed by deepfakes. It establishes fundamental rights and principles and addresses postmortem protections and freedom of expression thoughtfully. However, further detail on enforcement, intermediary responsibility, and guidelines for permissible uses could enhance its applicability and effectiveness.