When you purchase through links on our site, we may earn an affiliate commission.Heres how it works.
AI models have rapidly become a daily part of life for many of us.
How is your data being used for training?
Each AI model is different when it comes to its training process.
For example, it is common for models to analyze the conversations and history we have with chatbots.
This way, they can learn natural language, understanding, and generation.
This helps them to become more engaging in a human-like manner.
AI models are inherently complicated with billions of data points and millions of users.
ChatGPT
ChatGPT, arguably the best-known of the AI generators, makes this process easy for you.
From here, scroll down to data controls and change the setting improve the model for everyone to off.
This will stop ChatGPT from using your conversations and data for any model training.
Google Gemini
Opting out of Geminis data training is quite a simple process.
First, sign on to either the app or the desktop platform.
On desktop, go tothe Google Gemini dashboardand hit the activity button at the bottom left of the screen.
Next, turn off the Gemini apps activity setting.
This will stop Gemini from saving your activities and using them.
On theGoogle Geminiapp, you will need to click your account on the top right.
From here, you follow the same process, going into Gemini apps activity and turning the setting off.
If you are on Google Workspace for Gemini, youll need to be an admin to change this setting.
Anthropic Claude
Claude, by default, will not use your data for training purposes.
However, there are a few situations in which this will be different.
If your conversations are flagged to the trust and safety team, they might be used for training purposes.
This can happen if you break Claudes terms and conditions in your inputs and requests.
It is not entirely clear how your data is used outside of training purposes.
you could readAnthropics full privacy agreementto find out more.
Deepseek also automatically collects information including equipment models, IP addresses, system language, and equipment IDs.
This information can then be used by Deepseek to train its model.
Midjourney
Midjourneyis slightly more complicated due to the way it handles image data.
However, it is not clear if the images you generate are used for training data.
Midjourney also publishes images generated to its website, where users can rank them.
On these plans, you’re able to also make previously made images private.
GrokAI
WhileGrokdoes use your data for training, it is very easy to opt-out.
How you do this depends on how youre using Grok.
For X, go to your controls and then privacy and safety.
Here, you’ve got the option to also turn off any data sharing from X in general.
Adobe
Adobe is quite clear about its data training.
To opt out, go into the Adobe account privacy tweaks.
From here, enter the content analysis for product improvement section and toggle the setting to off.
This will completely opt-out your data and the things you make via Adobe being used for training.
However, they do make opting out of data retention fairly easy.
This will opt you out of any training process Grammarly uses.
