Disappear
from LLMs.

Legal assistance to request removal or correction of your personal data in chatbot responses and challenge its use. With lawyer Antonio Cattarini.

Assess my request

When a chatbot talks about you

A response may link your name to false information, outdated events or private details. Shield starts with what appears about you and why you want it removed.

False information

A chatbot attributes inaccurate facts, roles or statements to you. We document the response and assess a request for correction or removal.

Outdated events

A response repeats a past event without reflecting subsequent developments. We assess relevance, context and public interest before preparing the request.

Exposed personal data

Contact details, health information or other private details appear. We identify the data and assess grounds to request removal or restrict its use.

Responses, sources and data use

A generated response, a web page cited by a chatbot and the use of data for training are separate issues. We identify where to act and who should receive each request.

The aim is to ask that the disputed data no longer appear in responses from the relevant service. Where a chatbot retrieves it from an external source, we also assess that source within the agreed scope.

Your removal request

01. Document

We collect relevant examples with dates, questions, responses and any sources. We define which data to challenge and what outcome to request.

02. Identify the remedy

We identify controllers and relevant channels. We assess erasure, rectification, restriction or objection against the legal basis and circumstances of the case.

03. Prepare the request

We prepare a reasoned, documented request. Submission, representation and any identity evidence are agreed after the engagement is established.

04. Review the response

We assess the recipient’s response and document the outcome. Complaints, litigation and additional recipients require separate assessment and engagement unless already included.

An outcome to verify

The work produces targeted requests, a record of responses and an assessment of next steps. Removal from one service must be verified within that scope: it does not establish erasure from every model’s weights, the internet or all other chatbots.

Controllers normally respond within one month. Where permitted, they may extend by two further months, notifying you and explaining the extension within the first month. This is not a promise that the service will be completed within that period.

The rights behind your request

Rectification, erasure, restriction and objection have different conditions. The right to be forgotten is not absolute: the assessment also considers freedom of expression, public interest and other applicable lawful grounds.

OpenAI, for example, accepts requests to remove personal data from ChatGPT responses. These are assessed case by case and do not remove data from external websites or search engines.

Data subject rights: EDPB · Removal from ChatGPT responses: OpenAI

One engagement, one payment

The service uses a one-time professional fee for an agreed scope. Before payment, you receive a written proposal stating recipients, included activities, estimated timing and fee. It is not a subscription and has no automatic renewal. This page does not activate purchases or set a price.

You may exercise your rights yourself without engaging a lawyer. The fee pays for professional assistance; it is not a charge owed to the provider for exercising the right. Data-subject requests are generally free, subject to the exceptions in Article 12 GDPR.

Start with what AI says about you

Name the chatbot and briefly describe what you want removed or corrected. Do not attach identity documents, case files or mappings to the first message. We then agree suitable channels for the necessary information.