SynthID: Google's New Tool to Identify AI-Generated Images

The rise of artificial intelligence (AI) has led to the development of new tools that can create realistic images from scratch. These tools have the potential to be used for good, such as creating art or generating educational content. However, they can also be used for bad, such as creating deepfakes or spreading disinformation.



A deepfake is a type of synthetic media in which a person in an existing image or video is replaced with someone else's likeness. This can be done using AI to manipulate the pixels of the image or video. Deepfakes have been used to create fake news articles, political propaganda, and even revenge porn.

To combat the spread of deepfakes and other forms of disinformation, Google has developed a new tool called SynthID. SynthID is a digital watermark that is embedded into AI-generated images. The watermark is imperceptible to the human eye, but it can be detected by computers.

SynthID works by using a technique called steganography. Steganography is the practice of hiding information in a way that makes it difficult or impossible to detect. In the case of SynthID, the watermark is hidden in the noise of the image.

To use SynthID, you simply upload the image you want to test to the SynthID website. SynthID will then analyze the image and determine if it is AI-generated. If the image is AI-generated, SynthID will also be able to identify the AI model that was used to create it.

SynthID is a valuable tool for combating the spread of deepfakes and other forms of disinformation. However, it is important to remember that it is not perfect and cannot identify all AI-generated images.

Here are some additional things to keep in mind about SynthID:

  • SynthID is still under development, and its accuracy is not yet perfect.
  • SynthID can only identify AI-generated images that have been created using a specific set of AI models.
  • SynthID cannot identify images that have been edited or manipulated after they were created.

Despite these limitations, SynthID is a promising new tool that has the potential to help us combat the spread of deepfakes and other forms of disinformation.

In addition to the information above, here are some other things you can do to help identify AI-generated images:

  • Look for inconsistencies in the image. For example, if the lighting or shadows in the image are inconsistent, it may be AI-generated.
  • Look for unnatural-looking textures or objects in the image. AI-generated images often have unrealistic textures or objects.
  • Use a reverse image search to see if the image has been used elsewhere. If the image has been used elsewhere, it is more likely to be real.

By being aware of the signs of AI-generated images, you can help to protect yourself from being misled by disinformation.