During eDiscovery, images are often a critical piece of evidence but with the use of so called artificial intelligence programming it has become difficult to determine if a photograph has been generated or altered.
This video will help you in your ascertaining whether a photograph or video is an AI generated.
https://www.youtube.com/watch?v=q5_PrTvNypY&t=561s
Detecting AI in images involves looking for subtle inconsistencies like illogical shadows, impossible physics, repeating textures, weird hands/teeth, or garbled text, alongside using dedicated AI detection tools that analyze metadata or digital fingerprints, and performing reverse image searches to find the source. While visual inspection catches common errors, specialized software or metadata analysis offers deeper insights, though no single method is foolproof due to evolving AI capabilities.
Software that can help you detect AI generated images and data:
Grammarly
Suggestions from Microsoft:
In an ideal world, those posting online will adhere to ethical standards and disclose that their images were created by AI—and in fact, many social media platforms are working on ways to detect and generate these disclaimers automatically.
If you’re asking, “Is this image AI?” there are a few techniques that you can use to avoid getting duped or manipulated:
AI-generated images can still have limitations, especially when pertaining to fine details. Look for extra hands, fingers, or limbs in a photo of people, for example, or blurred faces in both people and animals. Close attention to details and distortions is essential to AI image detection.
Some generative AI tools automatically generate a watermark in the lower corner of an image, faintly in the background, or invisibly embedded within the image’s code. Detect AI images by checking these three areas.
Online images often include metadata such as when they were created, their file format, and the tools used to create or modify them. Digital photographs may also include additional information such as the make and model of the camera used and the camera settings. This information helps identify the image’s original source. Many AI images provide metadata that may even list the AI tool used—another great way to tell if an image is AI-generated.
Similarly, AI detection tools can extract watermarks and metadata from uploaded images to determine whether a piece of media is from a genuine source or created by a prompt.
As AI-generated images proliferate across the internet, it’s not always easy to gauge the intentions behind their human creators. This is why it’s so important to know how to tell if a photo is AI-generated. Avoid the potential for disinformation that AI-generated content might create while still using the power of everyday AI tools to fit your own needs.