top of page

DiscoG X Workshop 3: Human or AI?

  • May 23
  • 4 min read

Updated: 3 days ago

DiscoG X Workshop 3 banner celebrating 10 years of DiscoG Coding Academy, 23 May 2026, with Iman's Detection Game interface in foreground
Workshop 3 of DiscoG X, 23 May 2026. Ten workshops marking ten years of DiscoG Coding Academy.

Look at the painting below... Is it real, or did a machine make it? That question sounds straightforward. It is not. Workshop 3 of DiscoG X asked students to think carefully about where the line is, and what it means that the line is getting harder to find.


AI-generated painting of ballet dancers in pastel tutus rehearsing in an ornate theatre interior
AI-generated. A written description of style and subject matter was given to an image generation tool. No artist name or original artwork was used.


Meet Iman


Iman Idris Ahmed is a designer and developer. She recently graduated with a masters degree in Computer Science from UAL's Creative Computing Institute, where her research focused on how computers understand images and how we can make AI safe and explainable.


For her masters project, Iman built The Detection Game. A web-based game that presents students with a single image and asked whether it is AI-generated or human-made, why they think so, and how confident they are. Alongside the game, Iman used and AI model called D-HAID Detector, trained on over 100,000 images to detect the invisible frequency fingerprints inside images and analyse their meaning. When she compared the results, 84 human participants correctly identified AI-generated art 67.7% of the time. Her AI model achieved 88.6% accuracy.



Iman's Detection Game web interface showing a painting with Human-made and AI-Generated buttons and a confidence slider
Iman's Detection Game, Detecting the Diffused. Players decide whether each image is human-made or AI-generated and rate how confident they are.


We met Iman at her degree show and invited her to come and bring the question to our students, and co-host Workshop 3 "Human or AI?"


Human or AI was Workshop 3 / 10. You can see the full series here.



Adapting Iman's project for the workshop

Before the workshop, DiscoG built its own small version of Iman's game in Python using a Tkinter interface.


We started by putting together a bank of 40 paintings. All of the human-made works were sourced from Wikipedia Commons, old enough to be in the public domain. For the AI-generated paintings, we wrote prompt descriptions of what we wanted, feeding them into ChatGPT and Gemini. We did not upload any original artwork or name any specific artists. We wrote descriptions of style, subject matter and mood, and let the tools interpret them.



ChatGPT conversation showing a written style description prompt and the resulting AI-generated pointillist painting of a Parisian boulevard
One of the prompts used to generate AI paintings for the DiscoG game. A description of style, subject and period only. No artist names used.


The results were interesting. Given the descriptions alone, the tools made choices about composition and colour that sometimes feel unmistakably machine-made, and sometimes genuinely did not.


Below is an example using ChatGPT where it was able to recreate "Girl with a Pearl Earring" by Johannes Vermeer, from a simple prompt that only referenced the style.



ChatGPT conversation showing a Dutch Golden Age portrait description prompt and the resulting AI-generated painting of a young woman with a blue headwrap and pearl earring
Another prompt from the DiscoG image generation session. A description of 17th century Dutch portraiture style produced this result.


Our version of the game focused on the human guessing side only. A player is shown a painting, told whether it is human-made or AI-generated, and asked to say how confident they were in their guess.



DiscoG Human or AI Python tkinter game showing Van Gogh's The Starry Night with Human and AI buttons and a confidence percentage slider
The DiscoG version of the guessing game in play. Students chose Human or AI, then set their confidence level before checking the answer.


Workshop Stage 1 – Learning about Iman's Project

Iman opened the session by introducing the concept of AI-generated imagery, including deepfakes, and why being able to tell the difference matters. She covered misinformation, scams, copyright and how AI detects AI, and more.



DiscoG X Workshop 3 presentation slide asking why we need to detect AI-generated content, listing misinformation, scams, bullying and copyrights
From the Workshop 3 presentation. The reasons detecting AI-generated content matters go well beyond art.


Students played her original Detection Game, going head to head with images and debating their answers.



Workshop Stage 2 – Coding & their game

In the workshop, students added a key piece of logic to the game. The wrote the code that asks players how many images they want to see, then picks that many at random from the image bank. Working through it step by step, each student typed the logic, while more advanced coders added their own validation rules for user inputs.


Then they all played it and checked their results!



DiscoG Human or AI Python tkinter game results screen showing score, confidence rating and a Why this matters section
The results screen from the DiscoG version of the game, built in Python. Players see their score, their average confidence, and why being able to spot AI-generated art matters.


Why this matters

Iman's research sits at the intersection of art, technology and ethics, and asks a question that matters far beyond a game. If humans can only correctly spot AI-generated art 67.7% of the time, what does that mean for how we trust what we see online? What does it mean for artists whose style can be imitated by a tool in seconds?


These are questions that our students in Years 5 to 9 spent an afternoon thinking about, and they are questions that tech industry has not finished answering yet.



Bonus content:

Check out a few our favourite AI-generated images:





Join us this summer

DiscoG X is a short series of workshops for Years 5 to 9, but DiscoG Coding Academy runs weekly term-time classes during the academic year as well as Summer Holiday Bootcamps for students in Years 1 to 13.


Whatever stage your child is at, there is a place for them here.


Summer Bootcamps are intensive and hands-on, with Specialised Courses built around the same approach you've seen in this post. Spaces fill up quickly.



Not able to make summer? Register your interest for September term-time classes to be one of the first to hear when spaces open.


Any questions? Get in touch!


Comments


Commenting on this post isn't available anymore. Contact the site owner for more info.
bottom of page