"Have I Been Trained?" is a search tool that helps you find out if AI developers have used your data to train such systems. Similar to sites like "Have I Been Pwned" help you find out if your passwords have been leaked, "Have I Been Trained?" tries to do the same, but with AI training.
It’s easy to know when an AI has used your data to train itself when that information pops up in the news. Meta, for example, uses Instagram to train its AI, as does Salesforce with Slack. But many other companies train AI engines, and it’s not always easy to find out if they’re taking your data or work.
This page, in particular, helps you find out if any of your images are used in public datasets, which are repositories available to anyone for training AI models.
How to Use "Have I Been Trained?"
First, go to the website haveibeentrained.com. With the platform, you can do three different things:
- Upload a specific image: This will find similar and identical images and compare them to the public datasets used to train the models.
- Enter a website’s domain: If you enter a website’s URL, you will find the public images on it contained in a dataset. This search includes the domain and subdomains.
- Text search: Type in a term to find images tagged with words closely related to the text you entered.
Choose one of these three methods mentioned above—upload an image, type in keywords, or enter a URL—to carry out your search.
After you perform your search, go to the results page. You'll see the photos used in the datasets that match your search. Below each photo, you will see a link to the same image within the dataset.
Image | ‘Have I Been Trained?’
View 0 comments