The Digital Likeness Directive Lets You Tell AI: Don’t Recreate Me

Consider dying and then having a stranger build a chatbot that looks, sounds, and “thinks” like you one year later – without any permission from your family members whatsoever. Until this week, there has been no standard way to prevent that from happening. Perhaps, there might be.

On July 23, 2026, the company Authentic Interactions launched the Digital Likeness Directive (DLD), a voluntary open standard which will give individuals total control over the possibility of AI systems to recreate their appearance, voice, or persona. The draft of the directive is co-created with the help of StoryFile, Lookalike.com, and Authentex companies and is available for public comments until September 30, 2026.

The reason why the Digital Likeness Directive exists

In recent years, AI-generated recreations of real persons have gained popularity – from voice clones to deepfake videos to “griefbots” generated using the information about individuals gathered during their lifetime. However, the legal frameworks surrounding these technologies vary across countries and states – what is considered illegal content in one place might be perfectly legal somewhere else. On top of that, each of the platforms handles takedown requests related to recreations in its own (or non-existent) way. The Digital Likeness Directive, according to the company, can fix this problem: it won’t be a legal framework, but a technical and procedural standard to which any platform can adhere voluntarily.

How the Digital Likeness Directive works

The framework consists of two key elements.

• The Personal Directive – a document created by a living individual specifying what kind of recreations of them is allowed. It has six areas to regulate: formats (text, voice, or video); whether AI system is allowed only to replay recordings or is allowed to generate something completely new; who has access to the creation; commercial use; the source material which may be used to generate an AI-based representation of an individual; topics which can be discussed. The individual is also able to choose not to delve into all these details and simply declare that they don’t want any of their AI recreations at all.
• The Takedown Procedure – if an individual is concerned about the creation of some AI-based likeness, any platform adhering to the standard is bound to handle the issue within 14 days and provide valid proof of consents given or remove the content altogether. The platform is not involved in making any decisions – it is responsible only for checking whether there is any authorization.

What happens if you never register a directive

This is when the standard gets an interesting approach. If an individual didn’t bother to register their directives while they were alive, decision-making authority will pass to their family, based on the same legal principle as the decision-making regarding remains of the deceased person. That is, the company borrowed a decision-making model from end-of-life law and applied it to end-of-digital-life law in order to help families not to argue with a team of platform’s customer service.

Privacy built into the registry

Declining participation in the directive costs nothing and only involves identification – no ongoing relationships with Authentic Interactions are required. Notably, the registry doesn’t store any photos, voice samples, or biometrics, and can’t be accessed using facial and voice recognition. It is a deliberate decision – a database which is supposed to protect individuals from misuse should not become an object of such misuse itself.

Regulatory Background

As you can see, the timing of the announcement is not random. The federal NO FAKES Act passed the Senate Judiciary Committee, and a new set of European Union rules on synthetic media is going to be effective from August 2, 2026. According to Authentic Interactions, such laws mostly provide a solution to the problems after the fact – punish the violation, but don’t offer any practical tool for handling the consent of the user while the creation of a reproduction. And here is a standard which tries to do exactly that.

This Is a “Do Not Resurrect Me” Button

Most stories about this announcement will present it as yet another AI-consent policy. However, here is the angle worth paying attention to: the Digital Likeness Directive is the first attempt to give people a chance to pre-veto a creation of an AI-based recreation of themselves which could happen even decades after their death. It is a different request from the existing GDPR data deletion or platform-level deepfake reporting which deals with the content that already exists. The current standard attempts to handle the content that doesn’t exist yet.

The standard also silently raises an old question – who makes decisions about the body of the deceased? – in a brand new context – who decides about AI-based recreation of the deceased? The family law already knows how to handle such issues as organ donation and disposal of the remains of a deceased person. The Digital Likeness Directive is the first serious attempt to transfer this chain of decision-making to AI clones.

TheTweaks verdict

The Digital Likeness Directive is a clever, timely suggestion – but it is worth remembering what it really is: a voluntary standard offered by a company whose products (StoryFile, Lookalike.com) are based exactly on this AI recreation technology. It doesn’t mean that the standard is insincere, but means that its power lies in the adoption outside of the ecosystem of this company.

As long as big platforms like Meta, Google, or OpenAI don’t adhere to the standard, it is rather an industry-shaping proposition than an enforceable right. Nevertheless, as a first attempt to provide a free, verifiable, privacy-preserving way to say “don’t recreate me” while the regulators will not agree on the binding regulation for the next several years, it is an interesting move.

For more AI and tech news, visit TheTweaks AI News or check out our latest tech launches and big tech coverage.