Meta – Requiem for a Dream
Algorithms, Glitch Art and Invisible Censorship
Introduction
As artists, many of us rely on social media platforms such as Instagram, Facebook, Threads, tumblr or in the case of NFT’s X . These platforms allow us to share work, build reputations, connect with peers, exchange ideas and opportunities, and make our work visible to audiences and curators.
For glitch artists this dependence is even greater. Glitch art exists primarily online. It spreads through networks, communities and platforms. Without those networks it becomes far more difficult for the work to circulate, be discussed and develop.
However, these platforms are not neutral spaces. They are owned by corporations, governed by rules, and increasingly moderated by algorithms.
Algorithms as Gatekeepers
Most social media moderation is now automated. Algorithms are used to determine what content is recommended, what content is promoted, and what content is removed.
When these systems work incorrectly they can have significant consequences. Content can be removed, accounts can be restricted, visibility can be reduced, and users may find themselves effectively hidden from audiences without fully understanding why.
This process is often described as shadow banning: the reduction or suppression of visibility without an obvious ban. A user can continue posting while their reach and impact are quietly diminished.
In this environment the algorithm becomes a gatekeeper.
Community Standards and Cultural Bias
Meta describes its community standards as a way of keeping users safe and creating a welcoming environment. These standards are presented as universal, but they are inevitably shaped by cultural assumptions, political pressures and corporate priorities.
The standards are also subject to change.
For artists working internationally this raises important questions. Which cultural values are being enforced? Who decides what is acceptable? And what happens when an automated system misunderstands artistic work?
A Case Study: Glitch Art Misidentified as Pornography
In October 2024, glitch artist Maelle Lencot posted a piece of work in the Glitch Artists Collective Facebook group.
The work was identified by Meta's systems as violating rules relating to nudity and sexual activity.
The image contained no pornography.
Nevertheless, the post was removed and sanctions were applied. More importantly, the action threatened the status of the Glitch Artists Collective itself, a community of almost 100,000 members and one of the most important spaces for sharing glitch art online.
The incident illustrates a central problem: when an algorithm makes a mistake, the consequences extend beyond a single artwork.
Reviews and Contradictions
Meta provides an appeals process.
Eventually, after review, Maelle's work was restored. Meta acknowledged that the removal had been incorrect.
However, despite restoring the content, restrictions affecting the group remained in place.
This reveals a contradiction. The system recognised the original decision was wrong, yet the consequences of that decision continued.
The left hand did not appear to know what the right hand was doing.
Reduced Distribution and Content Demotion
Meta openly acknowledges that it reduces the distribution of certain content.
Content can be demoted, shown less frequently, or recommended less often. In practical terms this means fewer people see the work.
For artists, visibility is everything.
A work does not need to be removed to be censored. It only needs to become difficult to find.
When distribution is reduced, the effect can resemble censorship even when the content technically remains online.
Learning Systems and Feedback Loops
Algorithms learn from data and previous decisions.
This raises an important concern. If glitch art is repeatedly misidentified as problematic content, the system may reinforce its own mistakes.
Each incorrect classification becomes additional evidence for future classifications.
The result can be a feedback loop in which glitch art becomes increasingly likely to be flagged, restricted or removed.
In effect, glitch art may be exposing weaknesses within the moderation system itself.
Personal Experience
Following the removal of Maelle's work, I began paying closer attention to my own interactions with Meta's platforms.
A non-glitch image I posted on Instagram was removed for allegedly containing nudity and sexual activity. The same image was not removed from Threads.
Subsequently I noticed increasing numbers of glitch works being flagged, hidden or subjected to reduced visibility.
Friends and colleagues reported similar experiences. Comparable work posted elsewhere, such as Tumblr, often did not produce the same results.
While much of this evidence remains anecdotal, the pattern became increasingly difficult to ignore.
Self-Censorship
Perhaps the most significant consequence is not content removal but behavioural change.
When artists become aware that certain kinds of work may trigger moderation systems, they begin adjusting their behaviour.
They may alter images, avoid particular themes, or stop posting certain work altogether.
Fear of losing visibility becomes a powerful incentive to self-censor.
The algorithm does not need to explicitly ban artists. It merely needs to encourage them to police themselves.
The Glitch as Environment
Traditionally, glitch artists explore failures in systems.
We investigate the edges of software and hardware, searching for moments where systems reveal themselves through error.
In this case, however, the glitch is no longer confined to the artwork.
The glitch has become the environment in which the artwork circulates.
The moderation system itself becomes the site of investigation.
Beyond Art
These concerns extend beyond social media.
Similar forms of automated decision-making are increasingly used in areas such as facial recognition, policing, welfare administration and surveillance.
If systems cannot reliably distinguish glitch art from pornography, we should ask how reliable they are when making decisions with far greater consequences.
The issue is not simply technical failure. It is the social and political power granted to opaque systems.
Conclusion
Algorithms are not neutral.
They reflect the assumptions, priorities and biases of those who design and deploy them.
Meta's moderation systems shape what can be seen, what can be shared, and ultimately what kinds of cultural production are encouraged or discouraged.
For glitch artists this presents a paradox.
We work with error, disruption and uncertainty. Yet we increasingly depend on systems that struggle to recognize those qualities as legitimate artistic practice.
The question is not simply whether glitch art has glitched Meta's algorithms.
The larger question is what happens when algorithmic mistakes become part of the cultural infrastructure through which art is produced, distributed and understood.
And what is lost when visibility itself is controlled by systems that cannot explain their own decisions?
Final question
How do we as Europeans foster our own values online in social media spaces when those very values are not reflected in largely American owned and allied spaces with their own political agenda using an algorithm which promotes division, hatred and disinformation under the guise of free speech?
