No one can guarantee 100% protection for anything in life. 

You may have an image that is AI generated but our tool has returned a human-created result, or vice-versa. 

This is because AI companies are continuously improving the tools they use to generate AI media, and bad actors are always looking for ways around deepfake-detection algorithms. Much like anti-virus tools, this is an ever-evolving arms race. 

The models we use are constantly re-trained on new AI-generated content to keep ahead of criminals.

Yes, you can use VerifyLabs.AI Deepfake Detector anywhere in the world where you have an internet connection and web browser.

Having trouble using it? We suggest using a VPN as certain countries may wish to block the tool.

VerifyLabs.AI Deepfake Detector currently operates at up to 98% accuracy.

You can download our browser extension and iOS app here: verifylabs.ai

Simply click on the ‘Add credits‘ link in the app or extension, or log into your account on VerifyLabs.AI.

Audio: Right-click anywhere on the webpage and click “Verify audio”.

Image: right-click on your chosen image and click “Verify image”. If you are having issues, please check the image type is .webp, . jpg or .png. 

Video: right-click on your chosen image and choose “Verify Video”.

When I right-click on a video, I can’t see the “VerifyLabs.AI” logo. What should I do? Without closing the first menu, simply move the mouse to another part of the video and click again. The “Verify Video” option should be now visible.

How to check files on my PC/Mac: please click on the browser extension logo and then click on the ‘Upload button’ and choose your desired file from your computer.

• Download the app from your app store. 
• Open the app, and click on the three bars in the top left corner.
• Go to Settings and enter the credentials you chose when you signed up for your trial or your subscription. 
• You can use two methods for testing:

App method: for testing files that you have on your phone

• Open the app on your phone
• Press the ‘three bar’ menu in the left.
• Press ‘Media’ and choose the file you want to test.

‘Share’ method: this verifies images and audio directly from other apps using the iOS Share menu. 

Here’s how to set it up and use it:

• Find an image in your Photos or an audio file (like a voicemail).
• Tap the Share icon (a square with an arrow pointing up).
• In the row of apps, scroll all the way to the right and tap the More button (three dots).
• Tap Edit in the top-right corner.
• Scroll down to the ‘Suggestions‘ list, find VerifyLabs, and tap the ‘green +‘ icon to add it.
• To make it easier to access, find VerifyLabs in the ‘Favourites‘ list and drag it to the top.
• Tap Done. Our app will now appear at the front of your share options.

• Now, from any app, tap the Share icon on an image or audio file.
• Select VerifyLabs from the list of apps. This will send the file directly to the app for verification.

VerifyLabs.AI uses several sophisticated neural network architectures which scrutinise digital media in different ways. These networks are trained on vast datasets of both real and synthetic content, allowing them to learn subtle patterns and anomalies indicative of manipulation. They are adept at identifying inconsistencies in pixel distribution and facial features within images and videos, and excel at detecting temporal irregularities in audio and video sequences, such as unnatural lip movements or voice modulation. By analysing these minute discrepancies, our systems provide a probabilistic assessment of the media’s authenticity, highlighting potential deepfake indicators with remarkable accuracy.

In many countries, owning such materials is illegal.

Please work within the laws and government rules in your jurisdiction.

We do not store any images/video/audio and so you may use the tool to check them as long as you have the legal right to do so.

We’re glad you asked this question. Some websites or apps (eg: Instagram) do not actually share a photo or video when you click on ‘Share’. They try to share a URL instead. When this happens, the detector won’t know what to check and you may get errors.

To test images and video in these circumstances, it’s best to take a screenshot of an image and test is using the ‘App method’ (see above), or to use your phone’s screen recording function to record the chosen video clip to your photo library and then also use the ‘App method’ to test the video.

Important: for optimum results, make sure your image or video has no black bands (or other non-image elements) above, below or to the sides as this can lower results accuracy. Use the crop function to make sure you test only the parts of the image or video you need to verify.

If VerifyLabs.AI Deepfake Detector verifies a deepfake, you can choose your course of action.

This includes alerting the police or law enforcement authorities in your country. If you found the deepfake content on a social media platform, you could report it directly to them if you think it’s misleading or harmful.

The VerifyLabs.AI Deepfake Detector is a suite of tools built for desktop, mobile and API (Application Programming Interface). These tools are designed to detect whether or not an image, video, or piece of audio are generated by AI or are human-made in whole or in part.

Without closing the first menu, simply move the mouse to another part of the video and click again. The “Verify Video” option should be now visible.

Any image, video, or audio clip you send to us is processed in the cloud and then irretrievably deleted.

We do not store any content you send us for any reason.

We will record the fact that you did send us media for checking, and the result, but nothing else.

VerifyLabs.AI uses several sophisticated neural network architectures, and they are chosen based on the type and content of the media they are checking. Some are better at detecting AI-manipulated faces; others are better at detecting whether or not an entire image has been generated by AI.

Different neural nets will give you slightly different results. 

The different algorithms measure different characteristics of a piece of content, and some algorithms will give stronger results than others, depending on what tool made the original content.

Results for images or videos can change if they have been manipulated. For example, an image sent via WhatsApp is almost always processed to create a smaller file for easier transmission over the internet. This changes the characteristic of the file and can result in variations in scores. If you get a grey or red result, you’ll need to treat the image or other media with caution. 

COMING SOON: If you suspect that an image is a deepfake, but it still gives you a ‘Human Made’ result, you will be able to run it through a separate detector within the app for a deeper check.