Microsoft tests human ability to distinguish real and AI-created images

Last update: 05/08/2025
Author Isaac
  • Microsoft launches a global test to differentiate real images from those generated by IA.
  • 12.500 participants analyzed 287.000 images from a variety of sources.
  • The success rate barely reaches 62%-63%, bordering on chance.
  • Machines far outperform people in this visual challenge.

Microsoft AI Challenge Real Image

The lines between reality and artificiality seem to be blurring every day in today's digital environment. This phenomenon has prompted Microsoft to organize a global challenge , inviting thousands of internet users to test their ability to distinguish between photographs taken by real cameras and creations generated by artificial intelligence.

The experiment developed by Microsoft, accessible through the game 'Real or Not Quiz' , has established itself as one of the biggest online visual challenges of the moment. The system presents the user with various images—some authentic, others generated by AI—and asks them to decide in each case whether the photograph is real or synthetic. The goal: to explore the extent to which we trust our own visual judgment in the age of artificial intelligence.

A global challenge that reveals our limitations

Microsoft experiment distinguishes real image AI

The test mechanics are simple: participants view a selection of images and must classify them as real or AI-generated. In total, the database managed by Microsoft contains approximately 287.000 photographs , a mix of genuine images and content created using the most advanced visual generation engines (such as DALL-E 3, Midjourney v6, Stable Diffusion, Amazon Titan, and various GANs).

What's surprising is the users' poor accuracy . More than 12.500 people have completed the test, and the average success rate is only around 62% or 63% . This figure demonstrates that, currently, correctly identifying the origin of an image is not much easier than guessing it randomly . The quality of artificially created photos has reached a level that challenges even the most experienced users.

  Midjourney advanced commands: a complete guide to mastering AI

The results are collected and analyzed automatically, allowing each participant to compare their own performance with that of the other users. Simply visit https://www.realornotquiz.com/ to take on 15 images at a time and discover just how inaccurate our vision can be.

Why is it so hard for us to distinguish them?

The explanation goes far beyond simple generation errors. AI models have perfected their technique to the point of mimicking with remarkable accuracy the textures, lighting, and compositions typical of traditional photography . In fact, a significant portion of the successes achieved today are due to some users having learned to recognize recurring visual patterns—such as the type of blur, the color palette, or the sharpness—characteristic of certain image generators.

However, the task becomes more complicated if the generated image is based on a slightly modified authentic photograph or if it simulates "imperfections" typical of real cameras. In these cases, the success rate drops to 21-23% , reflecting our limited ability to unmask digital manipulation.

Some real-life photographs shown in the test are more misleading than the AI creations themselves. Images of military environments, unusual urban scenes, or extreme lighting tend to confuse participants the most, who tend to rate them as fake due to their unusual characteristics.

Humans versus machines: the advantage of algorithms

Alongside the user testing, Microsoft subjected the same set of images to its AI-generated content detector . The contrast is striking: while human participants rarely exceeded 63% accuracy, the detection system achieved rates exceeding 95% in any image category.

This reveals not only the gap between technological evolution and our perception, but also the growing importance of automated visual verification tools in a context where disinformation based on synthetic images is spreading rapidly. The problem is that these solutions are not yet accessible to everyone, nor are they available on most digital platforms.

  OpenAI is betting on an autonomous AI 'researcher' and accelerating its corporate transformation

Therefore, Microsoft insists on the need to introduce watermarks and robust verification systems so that users can identify the true origin of an image and not rely solely on their visual intuition.

What images deceive us the most?

The analysis reveals an interesting pattern: people tend to be more accurate with human portraits , possibly because our brains are trained to notice details and flaws in faces. However, identification fails spectacularly with landscapes, objects, and everyday scenes , where AI can more easily blend in. Furthermore, low-resolution images or those saved with generic names tend to go unnoticed as potential fakes.

Microsoft researchers themselves discovered that users who regularly use generative tools develop a certain "intuition" for recognizing patterns and textures. But, generally speaking, the technical quality of the new models means that errors are minimal and it's almost impossible to distinguish them without technological assistance.

Any internet user can take on the challenge and test their visual acuity by accessing Real or Not Quiz . Fifteen images are randomly selected from a vast database, and the results vary with each game. The goal is to determine if we can truly trust our eyes in the face of the unstoppable advance of artificial intelligence.

The experience is as educational as it is revealing: the line between reality and artificiality is increasingly blurred, and the need for visual literacy grows alongside the capabilities of generative systems. Only with new tools and greater awareness can we face the challenges of this new digital age, in which even a simple photo can shake our confidence.

Isaac

Isaac

Passionate writer about the world of bytes and technology in general. I love sharing my knowledge through writing, and that's what I'll do on this blog, show you all the most interesting things about gadgets, software, hardware, tech trends, and more. My goal is to help you navigate the digital world in a simple and entertaining way.