EchtNep Preparation

Proposal (Step 1)

Introduction

We would like to host a workshop on identifying AI-generated imagery. The rise of AI content is a painful development with great dangers for all layers of society. The lines between what is real and what is fake are becoming increasingly blurred as large language models are trained better and better, with more and more information. Still, fake images are not new and have been used to obfuscate reality for a long time.

The idea for the workshop is to bring examples of images, and asks participants to bring their own, and then determine whether they are legitimate or not. Together, we will share moments of being unsure whether something was AI or not, our feelings about that, and how we can address this in our communities. In small working groups, we will analyze the content and take collective notes. We would like to discuss potential consequences of ubiquitous AI imagery and how we can counteract doubts in our own ability to tell truths from fake.

Crowds & Places

The goal is to reach people from different backgrounds and generations. Ideally, we would like to host the event at Tante Nino in Rotterdam. Another option would be Leeszaal West (also in Rotterdam).

Technology & Social Practices

The workshop will largely revolve around discussion and shared experiences. We ask participants to bring their own images—either printed or on their phones or tablets. The images we provide would be printed, as this makes it easier to look at them all together. (Alternatively, we could use a computer and a big screen or beamer). We would also provide pens and paper to allow for note taking, scribbling and sharing visual ideas.

To make a collective zine out of our notes, we need a scanner and a computer. If we want to print copies during the workshop, we also need and a printer.

Another interesting practice could be to do research on the context of one specific image, and then share this with the group. This would require multiple devices with internet access.

Schedule

Timeframe: 3 hours

Introduction (30 min)

• Get everyone together. Introductions of the people who will be leading the workshop. Names of all participants.
• What is this afternoon about? A brief overview of the history of social media/the internet and how we have now entered the age of AI.

Break (10 min)

Group discussions (60 min)

• We get out the images we brought, and collect the content that has been brought by others.
• For every image, we go through the same process:

1. Speed round: Do you think this is AI or not? 
2. Describe the image: What do we see? Anything that that stands out? Ask the person who brought it about the context.
3. Pass around the printed cards and look for clues. Make notes, mark and highlight.
4. On bigger sheets of paper, we can note down general thoughts that might be unrelated to the specific image.
5. Ask again: Do you think this is AI or not?
6. Reveal a point from our list what to look out for.

Break (20 min)

Debriefing & creating a zine (60 min)

• What have we learned from the analysis? How can we read images in the future? We summarize our findings.
• We scan the mind maps, collect the (non)AI images, and digitize our notes. We use these to create a small zines that everyone in the workshop can take home.

Communication & Documentation

For now, the plan is to promote the event by printing posters and/or flyers and distributing them around the location of the workshop, as well as around similar communal spaces. We can also consider to post a sharepic in local Rotterdam groups on WhatsApp or Signal. Maybe there is a mailing newsletter of the venue?

For documenting, it would be very interesting to encourage participants to take notes (could be handwritten or on a device), then scan or photograph those and make a zine out of them—possibly right away so people can take it home if they want to. To make the information accessible to people engaging with the and around the location, we would like to share our take-aways on the website of the venue as a recap of the event. If possible in terms of time, it would be cool to make a website or a small publication that documents the workshop steps and our learnings in a more polished way.

Process

Pad

https://pad.xpub.nl/p/idai

Organising

02-02-2026: Reaching out to Tante Nino We start out by writing an e-mail. There's no response for quite a while, so Daan sets up a meeting with Maaike from Tante Nino via their website. Some questions that we gather beforehand:

• Which info do you need from us upfront?
• How do feel about hosting an event in English only? Are you doing that regularly, or would you prefer if we aim to promote it in Dutch as well, and offer resources and feedback in both languages?
• Do you have a scanner and/or printer available?
• Do you have a newsletter? Or other practices to inform people of events at Tante Nino?

13-02-2026: Meeting with Maaike & Willem at Tante Nino

Lea made a work-in-progress flyer. Needs to be updated with the new title!

Maaike relates to our experiences and shares her own anecdotes. She suggests that we work with the Dreamland101, a project from the local elementary school, and target the workshop to the kids aged 8-12. Her and Daan come up with a title idea: Echt Nep?! (meaning Really fake?! — an expression used frequently in Dutch). Possible dates for the workshop could be March 4 (Wed) or March 6 (Fri). Needed to go on:

• a flyer/sharepic
• a short text to promote the session
• decision on the date

01-03-2026: Reorientation The communication with the different actors is slow-moving: Maaike only works at Tante Nino two days a week, and we have no direct contact to Dreamlab010. March 4 (or 6) is approaching FAST, and there are no clear plans arranged. According to Maaike, it might not be feasible to host the workshop in March. We decide to look for other venues, while still being open to doing it at Tante Nino, possibly after the Special Issue.

02-03-2026: Reaching out to Leeszaal West We write an e-mail to Leeszaal West, and decide to go there in person the coming Wednesday. Doriane gives us some feedback and food for thought:

• Document the process of organizing and going through the loops!
• If you want participants to bring their own material to the workshop, be as specific as possible. Example: Gossipy questions from class. "When was the last time you experienced X?"
• How are you going to navigate different viewpoints on AI among the participants? AI is political, and people like to express their opinions (loudly). It is important to include a plurality of beliefs and emotions, but it might be good to clarify some common grounds / negative aspects of AI. In the promotional communication, we could already try to target people who are at least somewhat critical of it.

04-03-2026: Visiting Leeszaal West At 11:30 we met in front of the Leeszaal and went in there. Talked to the volunteers who were having coffee, said the person we needed was Dolf, was coming in a little bit. The person who told us this kindly asked us what the workshop we wanted to host was about. He said it would really fit the space and they'll probably have a spot for us. The atmosphere of the place is very calm and welcoming. I think it's really an actual community space. Lea went to work with her writing group and we (Ema and Daan) were told by Dolf that we needed Michiel for this. This was kinda funny as it shows to me the spontaneity and informal environment the place. Dolf gave us Michiel is number, we called him and he invited us to a meeting on Saturday morning. The comforted us on the phone by saying they're super flexible and they are likely to fit it in!

07-03-2026: Meeting with Michiel at Leeszaal West

New flyer!

• Via facebook and instagram of Leeszaal we can share the flyer and text.
• They can co-post it with us - maybe via XPUB instagram?
• We could do the workshop multiple times
• Maybe put a qr-code on the flyer so people can apply
• If it's gonna be 15 or more people participating, we might need to consider the back of the Leeszaal. That would be 50 euros for a day (?)
• We need to bring a printer ourselves if we want to print zines, but it would be okay to use it!
• Not on Saturday afternoons - Taalcafe NL (14:00 - 16:00)
• Not on Wednesday, 18 March - municipality election
• Not on Monday 23 March - XPUB2 presentations at Leeszaal
• At other times: free to walk in with a group up to 10 people
• Communicate with Leeszaal which day(s) and with how many people we want to be there, so the volunteers know what to expect.
• Contact information: verhuringen@leeszaalrotterdamwest.nl & promotie@leeszaalrotterdamwest.nl

To-Do's:

• Find dates we would like to host! Maybe 19th or 20th and 26th or 27th of March?
• Make updated flyers - for these dates
• Make final small text with these dates on it
• Share flyer and text via chatgroups & instagram.
• Print simple A5 flyer in b/w
• What printer can we bring?

Lists: What's important?

Dangers surrounding AI content

• Slop is everywhere: The internet as well as the real world are increasingly flooded with low-effort content, and it is hard to find something valuable in between.
• Creating realistic fakes is made easy: scams are harder to detects, identities get stolen, and women get sexualized without consents.
• Visual crafts (and others such as writing or coding) loose recognition.
• People are less inclined to actually learn a skill.

> Add examples

Common giveaways that something is AI-generated

• weird hands with more than five fingers
• faces that are too perfect / smooth
• glitches of things growing into each other)
• in videos or multiple images: moving doorframes, furniture moving around in the same room
• Logo's being mimicked, looking legit from a distance but actually its abstract shapes when looking closer
• Butchered text (look closely!)
• Multiple contexts juxtaposed to each other (Colosseum within the streets of Venice)
• Heavily saturated and bright colors & heavy contrast
• Super bright/focused main subject, everything in the back blurred
• Jagged patterns/structures

> Add examples!

Contexts!

• What is the image showing?
• Who published it? What do I know about the publisher?
• What is it trying to communicate? Is it just slop or trying to sell me a product or idea?
• Are there multiple examples depicting the same item or event?
• Do I know someone who has more experience with the thing in the picture or detailed understanding of the situation?

Things we need to bring to Leeszaal

• printer (with double sided printing) > Daan will ask Guus
• flatbed scanner > Daan has one
• A4 paper to print on > Look in the studio on Wednesday
• paper to take notes on > Look in the studio (or WdKa shop?) on Wednesday
• printed example images > Print in school on Wednesday
• printed lists of giveaways and contexts (each 3-4 times) > Print in school on Wednesday
• scissors, pens, glue > Bring from the studio and from home

Reading & Sources

In just the last few years, even months, the quality of AI content has improved immensely. By now, obvious six-finger glitches and horribly rendered texts are mostly a relict of the past, and there is no surefire way to tell if an image is generated or has been digitally altered. Online tools are supposed to help with identifying, but they are not reliable, and watermarks used by some chatbots can easily be removed. So—how to decide if an image is real or fake?

In this article, Meaghan Looram—director of photography at the New York Times—describes the process that journalists go through to determine the authenticity of a photo. She, too, states that AI-detection tools leave room for uncertainty. Instead of relying on them, the most important step is to look for context. Context can include the source of the image, or other depictions and reports of the same event. This is relevant for ones one understanding, as well as for sharing images further. Looram also mentions a back-and-forth discussion within the team. This might be good journalistic practice before publishing anything, but gossip can also be very helpful in personal moments of doubt. Lastly, Looram emphasizes the importance of making human judgment calls!

___

[Anna Lowenhaupt Tsing, author of The Mushroom at the End of the World,] positions alienation as a key mechanism of capitalist simplification and control. Alienation severs the interdependence and relational value of humans and more-than-humans, commodifying them and reducing them to exchangeable assets—a process she describes as "barcoding life".
- Ruby Bayley

___

AI is political! In that regard, the article Do Artifacts Have Politics? (Langdon Winner) is very interesting. We need to study inherit qualities of new technologies, and pay attention to the context in which they are emerging. Assessing the political consequences of a system (ideally—haha) needs to happen before it's widespread adoption, as the social and environmental requirements change (sometimes drastically) with the use of a technology. In the text, two types of political technologies are explained—1: A thing becomes a way of settling an issue in a particular community. And 2: A thing is inherently political. At least in their current form, LLMs are part of the second category. As Jürgen Geuter (tante) puts it in a response to an article published by Cory Doctorow: "Capable LLMs require a logic of dominance and of disregarding consent of the people producing the artifacts that are the raw material for the system. LLMs are based on extraction, exploitation and subjugation […]."

(How) Do we discuss that in the workshop?

___

links

https://www.theguardian.com/world/2026/jan/29/south-korea-world-first-ai-regulation-laws

https://www.cambridge.org/core/product/identifier/S0003055423001454/type/journal_article

https://www.nytimes.com/2026/01/04/insider/how-the-times-assessed-maduro-photos.html

https://www.cnn.com/2026/02/05/style/fake-exhibition-amsterdam-rijksmuseum-scli-intl

https://faculty.cc.gatech.edu/~beki/cs4001/Winner.pdf

https://tante.cc/2026/02/20/acting-ethical-in-an-imperfect-world/

https://www.youtube.com/watch?v=M4TXO4kQwSQ

Feedback from Doriane

Hey those are a few notes (for myself at the moment) but let's discuss all of those next monday!

in the intro you state "to obfuscate reality" as the main concerning thing about "the rise of AI content". i think this is def one aspect, but there are more danger to that: fake images have always existed and have a long history, yet there is something more specific about AI content. on the top of my head i'm thinking of things like

• zombification and slop - how AI content is fueling infinite feeds of things to consume where we genuinely don't care anymore if it's "true" or not but just want endless generation of empty dopamine-induced quick satisfaction
• another aspect is stealing and dis-forming other people's identity and sexualising them (mainly women), what is wrong isn't the "truthness" of such images - as it should be ok to purposefully sexualise yourself if that is your own will (anti-slutshaming) - but how AI (but also law and platform!) makes it easy for one person to steal and change the identity of another (?)
• it also impacted all practices of "visual craft" from drawing to graphic design, that are now seen irrelevant by some because they can "make the same with AI", in that case the impact is how it existence is in friction the cultural importance of intentional visual craft - and with the (loss of) valuable skills that goes with it)

it could be a very good exercise for you to list such things (and of course expend it depending on peoples stories) after the workshop! also i love this (the feeling, the doubt and creating solidarity and connection in them) - but try to do it, through note taking? through making lists (of feeling, of doubt) ? if you can look into some ref or an example for each of those in the news or in essays it would be amazing ;) if you want i could also forward you some essays or news on the topic.

Together, we will share moments of being unsure whether something was AI or not, our feelings about that, and how we can address this in our communities.

again i think that developing the skill of detecting when an image is AI or not is good for what you call "detecting the truth", but not only, it's also about detecting the "traps of attention and slop", or detecting "when someone is stealing someone else image", or when "the culture of visual craft is being attacked or not respected" - which all 3 aren't inherently linked to notions of truth (!)

for the images that people bring it could be good to ask them (of course why they picked that one - linked to a specific feeling ?), but also in which context did it existed and to document that too! (from insta, from tiktok, over vinted, in a DM, printed!) tracking the context in which people are most confronted to those also says something on how digital places are situated (for example there will be overlap of context, and some place - like the fediverse are i think less prone to that).

something else that matter for me to say, is that you are basically giving a collective therapy session over AI generated content (which i think is amazing! to be mostly computer-less and really focused on conversation), and I'm all for giving autonomy to individuals. However when approaching political global problems like that one they are 2 directions:

• individual gestures (being better at recognizing, and healing or understanding better our feelings)
• infrastructural actions (banning genAI over certain platform - like certain mastodon instance, or lobbying for a law that all social networks should have to respect, implementing genAI "waterwarks", etc)

i think for preparations, it would be good to research and list some "infrastructural actions" that have been taken already, either by small communities or by group lobbying for them in the law. My opinion is that both individual gestures and infrastructural actions goes together, but since your workshop focus on the first one, i still think it's good to do even a small research and have 2 or 3 initiatives that have been done in the other direction!

finally, i think that 2h is a bit optimistic. When we do our "gossip round" in class it usually take 1h ;) and personnally i wouldn't shorten it further. what about extending your workshop to 3h?