AI or not? new web tool detects artificial intelligence images in less than a second

Optic ‘Ai or not’ checks fabricated images

AI company Optic battles misinformation dispersed by AI-generated images with its web tool AI or Not. Its engine swiftly scans images and tells the users whether or not the image has been generated by artificial intelligence or by a human. The web tool contributes to checking deceptive content that may be caused by the use of AI image generators such as Midjourney, Stable Diffusion, DALL-E, and GAN. Optic claims that AI or Not is 95% precise with the result it gives the users through its ‘own state-of-the-art algorithms designed to provide highly accurate results.’

A concern that may come up can be related to privacy issues when users upload images for AI or Not to check. Optic states that all the uploaded images and inputted URLs are not stored on its servers ‘longer than necessary for analysis, and we adhere to industry best practices and relevant data protection regulations.’ It hints at the team still storing information but reiterates that it will be for analysis purposes. Such an attempt may help the team boost its dubbed advanced algorithms and machine learning techniques that analyze images and detect signs of Al generation.

images courtesy of Optic AI or Not (unless stated)

‘AI or Not’ by optic may help avoid image misinformation

In March 2023, designboom reported a series of viral made-up images circulating social media which included the Pope wearing a puffer jacket and Donald Trump’s violent arrest. While it has already been ruled out that these images are generated by AI, social media users might have still been lured into thinking that they were real (model and TV personality Chrissy Teigen had shared the same image on Twitter before she deleted it upon realizing it was a hoax).

There is a tendency that internet citizens might take the bait of AI-generated images especially if the images in question are close to subjects they resonate well with such as politics, entertainment, and culture. Web tools like Optic’s AI or Not might be able to avoid the spread of image misinformation as it compares the input image to known patterns and characteristics of various Al models and human-made images to determine the origin of the content. Users can also report to Optic if their result is true or false to help boost their checking system.

So far, AI or Not can only detect if the image is generated by artificial intelligence. In our recent coverage on the batteries of Nyobolt electric sports car, we ran a rendered image of the car on AI or Not. The web tool detects the image is generated by a human. Whether or not Optic will include render detection as part of its plan is still yet to be pinned down. In the meantime, it has hopped the wagon on verifying the truth behind the creation of photographs.

the image we tried is from Mateo Fernandez (read the full story here)