OpenAI and Privacy

Now that almost everybody is using ChatGPT or some form of AI in their lives and it is a known fact that all AI models train themselves with the data we give it. As we saw in another post, they also train themselves with the content on the net or scrape videos on YouTube amongst other things.

OpenAI, is the company that primarily deals with AI research and deployment. They primarily work with AI enabled technologies for the betterment of humanity. ChatGPT is their first child that was released in November 2022. ChatGPT took the technical world by storm and all wanted to taste the new AI enabled pie.

What could ChatGPT do? Well, ChatGPT could write code, do creative writing, give technical explanations, give reasons for various questions and more. Here is a detailed explanation from OpenAI’s site about what ChatGPT could do;

Source; https://openai.com/chatgpt

Unfortunately, with us asking so many questions, LLMs do understand us and our personal habits slowly. They do know our name, email address and what questions we ask. LLMs might form a digital persona of us unintentionally.

Large language models which train ChatGPT learn from three sources of information to train themselves. They are;

a. The big wide Internet

b. third party sources

c. information given by users like you and the trainers who are training the LLMs

Privacy and OpenAI;

Since our personal information and the questions we ask is used to train LLMs, privacy of users is at risk. With a lot of information and lot of questions, we are a goldmine of information for AI! What does OpenAI say about this? OpenAI does state that certain data is retained by OpenAI and steps are taken to reduce the personal information that is sent for training the model.

For example, if I login and ask ChatGPT a question such as ‘Where should I study?’

it will know that you do intend to study and can train its models along those lines with follow up answers.

How do we prevent this?

If we do not want ChatGPT to train itself with your information, you can click on the privacy portal of OpenAI and opt out of it.

Here are the steps for the same;

You can click on ‘Make a Privacy Request’ and request not to train on your content.

Once you have submitted your request, an email will be sent to confirm your request. Very soon after, you will get an email confirming that your data will not be used to train its models.

It is this small step that will ensure that your privacy is upheld on LLMs. It is also continuously necessary to keep track of all privacy updates and follow them accurately to remain safe online in the AI age!

This post is for alphabet ‘O’ for #BlogchatterA2Z 2024!