Fake or real? Here’s how to identify AI-generated images using ChatGPT and Gemini

AI-generated images are getting difficult to spot now. A few months back, it was easier to tell when a picture was fake because the faces looked odd, fingers appeared broken, or backgrounds made no sense. But newer AI tools have become much better, and that is exactly why more people are now getting confused between real and AI-made photos online.  Also Read: Why Google ‘disregarded’ the word ‘disregard’: Here’s what actually happened

From viral celebrity pictures to fake accident photos and political posters, AI-generated visuals are spreading across social media very quickly. Sometimes, users share them without even realising they are not real.  Also Read: You can now use ChatGPT directly inside PowerPoint to make presentations

Because of this, companies like Google and OpenAI are now adding tools that can help users verify images.  Also Read: Spotify wants fans to stay longer with AI podcasts, remixes and exclusive ticket access

What usually gives away an AI image 

Even though AI tools have improved a lot, they still make mistakes sometimes. One of the easiest ways to check an image is by zooming in and looking carefully at smaller details. 

Hands, fingers, earrings, glasses, reflections, teeth, and text written in the background can still look unusual in AI-generated images. In many cases, signboards contain random letters, while faces in crowds may look repeated or distorted. 

Lighting can also feel strange. Sometimes the shadows do not match properly, or the skin texture appears too smooth compared to a real camera photo. 

Another thing people often miss is context. If a shocking image suddenly goes viral but no trusted source is posting it, there is always a chance that it could be AI-generated or heavily edited. 

How to check AI images using ChatGPT tools 

OpenAI has started offering image verification support for AI-generated visuals created using its systems. 

Here’s how you can check an image: 

1. Open a browser on your device. 
2. Search for OpenAI’s image verification tool. 
3. Upload the image you want to verify. 
4. The system scans the picture for AI-related markers and credentials. 
5. Once the process is complete, it tells you whether the image appears to be generated or edited using OpenAI tools. 

The tool mainly checks for hidden identifiers added during image creation. 

Google Gemini can also help identify AI images 

Google Gemini is also getting image verification support through Google’s SynthID technology. 

To use it: 

1. Open the Gemini app on your phone. 
2. Upload the image from your gallery or files section. 
3. Ask Gemini questions like “Is this AI-generated?” 
4. The app will analyse the image and look for AI watermarks or editing patterns. 
5. Gemini may also explain whether parts of the image appear AI-generated. 

Google says SynthID watermarking has already been used across billions of AI-generated images. 

Reverse image search still helps 

Apart from AI detection tools, reverse image search still remains useful. You can upload an image to search engines and check where it first appeared online. If the same image keeps appearing with different stories or no proper source attached to it, that is usually a warning sign. 

Add Techlusive as a Preferred Source

AI-generated content is improving very quickly, which means spotting fake images manually is only going to get harder from here.