How-to·May 5, 2025, 10:48

Wrong birth date in ChatGPT triggers GDPR complaint against OpenAI

AI-generated from the sources listed.

According to the privacy organization noyb, OpenAI has refused to correct an incorrect birth date in ChatGPT. The organization has filed a complaint with the Austrian data protection authority.

AI-generated image

What do you do if ChatGPT says something wrong about you? That question has become pressing in a complaint from the European privacy organization noyb. It was filed on April 29, 2024, with the Austrian data protection authority, DSB.

Wrong birth date

The complaint concerns ChatGPT repeatedly giving an incorrect birth date for the complainant. According to noyb, OpenAI refused to correct or delete the information.

Noyb also alleges that OpenAI responded inadequately to an access request. This was a request to see what personal data is stored and where it comes from. According to noyb, this violates GDPR Articles 5, 15 and 16. GDPR is the EU's set of rules on the protection of personal data.

According to noyb, OpenAI has admitted that the company cannot correct incorrect information. OpenAI pointed out that accuracy in large language models is still an area of research. The company can only filter or block certain prompts (the questions users type into ChatGPT). It cannot selectively remove information about a specific person.

Noyb is demanding that DSB investigate OpenAI's data processing and order the company to comply with the access request. The processing must be brought into line with GDPR, and the organization also wants a fine imposed.

OpenAI's form

OpenAI has introduced a web form where people can ask to have personal data removed from ChatGPT and related products. According to OpenAI, submitting a request does not guarantee that the information will be removed. The company will assess it and weigh privacy and data protection against societal interests such as access to information.

It is technically difficult to correct or remove individual pieces of information in a trained language model. OpenAI has previously blocked prompts containing names, but has called it an unsustainable long-term solution because it is too cumbersome. The news site The Decoder assesses that the form only addresses symptoms and not the underlying problem, namely that language models make up information.

Advice from a consulting firm

The consulting firm Kalicube recommends creating your own main page about yourself, a so-called "Entity Home," and structuring your presence on credible sites. You should also give AI engines such as ChatGPT and Gemini clear data about yourself. Kalicube claims that this saves time spent explaining yourself and opens up business opportunities. That is the firm's own claim and a sales pitch.

The complaint and the form date from 2024. The outcome of the complaint is not stated in the sources, and the procedures may have changed since.

Sources

Found via The Next Web

More on this topic

Get the week's AI news in your inbox

Choose your level, topics and length. One email a week, unsubscribe at any time.

Subscribe to Promptly Newsletter