Putting aside the already exhausting hype around ChatGPT and large language models generating — besides lies — embarrassing poems and drier articles than a constipated asscrack, I wanted to quickly write an article for those and only those who haven't gotten high from far too much AI shrooms, and prefer keeping their heads on their shoulders and their feet on the ground. I repeat, this article is not for ChatGPT or OpenAI fans. If you are one of them, go, read something else, something that strokes your borderline delusional robot-obsessed ego. Sell your soul to AI, for all I care. 🤷‍♂️

This will possibly be one of the shortest ever articles I will write about tech or AI, which says a lot, as I am known for loving a good 6+ minute column. Time is of the utmost essence, however, so I decided that despite being sleep-deprived, I will offer my contribution to raising awareness around the possibility of removing your information from ChatGPT, DALL-E, or in general, OpenAI.

Why would you want to remove your information?

There are plenty of reasons, but after the Italian scandal that saw ChatGPT being banned temporarily, it became obvious that OpenAI isn't doing things by the book. In fact, the general business model of AI companies seems to be a "no book, no rules, no regulations" one, where, unless told otherwise, everything is fair game. Any person even partially sane would argue that such an approach is no better than Hitler's when it came to killing the Jews. There was no law specifically prohibiting killing Jews or sending them to labour camps, but apart from psychopaths, everyone will agree that killing people because of their ethnicity or race, is wrong. The same applies to AI, in particular LLMs. Just because the law didn't specifically state that personal data or data publicly available on the internet cannot be used by large language models, does not imply that it's fair game.

So, without further ado, let's see the reasons why you'd want to remove your data:

  • You are an EU citizen and according to GDPR — an otherwise very anal law, but particularly useful now — you have the right to be "forgotten", or your data not to be processed.
  • You are a citizen of a country that has other similar data protection laws, even if not as explicit as the EU's GDPR.
  • You don't want your personal information to end up in responses provided by the chatbot. Seems like common sense to me.
  • You want your conversations with ChatGPT to stay between you and the AI, never to be used for further training. After all, why would you want to help a company make money, using your conversations? That's not a symbiosis, it's shameless exploitation.
  • Your children used ChatGPT before parental tools and age limit were introduced, and you are worried about the information they may have shared. Let's be honest. Kids do have a tendency to overshare online.

How to remove your information?

There are two forms you must fill out. I'll be honest, I take issue with even having to fill these out, because these are both opt-out forms, which essentially means, by default you are opted in, and while that may not (yet) be illegal, it's certainly not ethical as it assumes — aka, hopes — people are either ignorant or ill-informed, without taking any action to remediate that. How little they care about any of this, is also evident from the forms, one a hastily put together HubSpot form, the other one is even more basic, just a Google Docs form.

So, what are the two forms, and what do you ensure with each?

  • The User Content Opt Out Request form ensures no information you provide to ChatGPT or DALL-E from that moment on, will be used for further AI training. This is a very short, 3-fields form. It will take you but a minute to complete and submit.
  • The second one is the OpenAI Personal Data Removal Request form, and this is related to your rights under data protection laws, such as GDPR. This one is a lot longer and requires you to fill in up to 17 fields! It even asks you to provide proof that they already generate answers directly related to you. I, for one, find this despicable. OpenAI practically stole data from its current or future users, but they're still asking the user to do the legwork to prove it. Clearly, the form is created in a way to make people give up, and just let OpenAI do whatever the hell they feel like.

As promised, I kept this article pretty short, because I think it's crucial for every DALL-E and ChatGPT user to know they have rights, and they are not powerless in the face of AI. Scrutiny will only increase, and many countries are already working on laws that will radically change the way AI companies can operate and how much they can get away with. That's a good thing, as no technological evolution should be left to its own devices, unsupervised. Fill out the forms, take control over your data, and while you're at it, toggle chat history and training off on the site as well.

Attila Vago — Software Engineer improving the world one line of code at a time. Cool nerd since forever, writer of codes and blogs. Web accessibility advocate, LEGO fan, vinyl record collector. Loves craft beer! Read my Hello story here! Subscribe and/or become a member for more stories about LEGO, tech, coding and accessibility! For my less regular readers, I also write about random bits and writing.