Artificial intelligence has not only introduced new tools into creative work; it has also created an entirely new vocabulary. Terms that once belonged to science fiction, online communities or technical discussions are now appearing in design studios, marketing teams, technology conferences and everyday conversations between creatives.
You do not necessarily need to start using every one of these expressions yourself. However, understanding them matters because they increasingly shape discussions about how AI is changing creative industries. Designers, developers, writers, marketers and decision-makers are using this language to describe new behaviours, frustrations and risks that did not really exist a few years ago.
1. Clanker
The word clanker originally comes from Star Wars, where it was used as a derogatory term for battle droids. Internet culture has since repurposed it as a dismissive nickname for frustrating AI systems and automated services.
You might hear someone use the term after spending 20 minutes trapped inside an unhelpful customer-service chatbot or repeatedly receiving irrelevant answers from an automated assistant. It captures a very specific kind of irritation: knowing that a human could probably resolve the problem quickly while the machine continues sending you around in circles.
The popularity of the term also reflects changing attitudes towards automation. People may appreciate AI when it works, but patience disappears quickly when an automated system becomes an obstacle rather than a convenience.
2. Necromarketing And Delebs
Necromarketing refers to using artificial intelligence and digital recreation technologies to bring deceased celebrities back into advertising or entertainment. The resulting virtual personalities are sometimes called delebs, short for digital celebrities.
A deceased actor, musician or public figure can theoretically appear in a completely new commercial despite having died decades earlier. AI-generated voices, synthetic video and increasingly sophisticated visual effects make these recreations far more convincing than traditional archival footage.
Technically, the possibilities are impressive. Ethically, things become much more complicated. A dead celebrity obviously cannot personally approve a new endorsement, raising questions about who has authority over their likeness, voice and reputation.
As the technology improves, disputes involving estates, advertisers, studios and personality rights are likely to become increasingly common.
3. AI Slop And Sloppers
Few recent AI terms have spread as quickly as AI slop. It describes the enormous volume of low-quality, disposable content being produced by generative AI and pushed onto websites, social networks and other digital platforms.
The obvious examples include strange AI images with malformed hands or nonsensical details, but slop extends well beyond pictures. It can include automatically generated articles with little original value, repetitive social media posts, meaningless videos and entire websites created primarily to capture search traffic.
People who operate these content-production systems are sometimes dismissively referred to as sloppers. Their strategy generally prioritises volume over craftsmanship, originality or usefulness.
Ironically, the rise of AI slop may ultimately make genuine creative work more valuable. When audiences become surrounded by generic synthetic material, thoughtful work made with clear human intent becomes easier to recognise.
4. AI Washing
Most people are already familiar with greenwashing, where companies exaggerate or misrepresent their environmental credentials. AI washing applies the same idea to artificial intelligence.
A company may describe a product as "AI-powered" even when the underlying functionality is relatively ordinary. A supposedly intelligent assistant might simply be a conventional search interface, while something marketed as machine learning could actually operate through a small collection of predefined rules.
The AI label has enormous marketing value at the moment. It attracts attention from customers, investors and the media, which naturally encourages companies to attach the terminology to almost everything.
The best response is simple: whenever a product claims to use AI, ask what the AI is actually doing. If nobody can explain the answer clearly, the branding may be more impressive than the technology.
5. AI Glazing
Glazing originally became popular as internet slang for excessively praising someone or something. In AI conversations, AI glazing describes enthusiasm that has moved far beyond reasonable optimism.
AI glazers tend to interpret almost every new feature as revolutionary. A modest productivity tool becomes evidence that an entire profession is about to disappear, while a slightly improved chatbot is presented as proof that human creativity has already become obsolete.
There is nothing wrong with being excited about technological progress. The problem appears when enthusiasm prevents meaningful discussion about limitations, costs or unintended consequences.
Creative professionals benefit from maintaining some scepticism. AI can be extremely useful without every new feature representing the beginning of a completely transformed civilisation.
6. Alignment Tax
Generative AI systems are typically surrounded by safety rules intended to prevent harmful, offensive or otherwise undesirable outputs. These protections exist for good reasons, particularly given how early language models could reproduce extremely problematic material from their training data.
The term alignment tax describes what users lose in exchange for those safeguards. Greater safety can sometimes mean less creative freedom, particularly when the model becomes overly cautious about anything controversial, provocative or unusual.
This becomes noticeable during brainstorming. Creative work frequently requires exploring ideas that may initially sound strange, uncomfortable or inappropriate before finding something interesting inside them. A model that constantly refuses to explore unconventional territory can produce ideas that feel safe but predictable.
The challenge is finding the right balance. Too little alignment creates obvious risks, while too much can make a creative assistant frustratingly bland.
7. AI Jailbreaking
The term jailbreaking previously referred mostly to modifying devices such as smartphones so they could run software or perform actions their manufacturers did not officially allow. With AI, the meaning has shifted towards bypassing a model's built-in restrictions.
Users may create elaborate role-playing scenarios, indirect instructions or complicated word games designed to convince the model to generate something it would normally refuse. Entire online communities experiment with discovering these weaknesses.
Sometimes the objective is relatively harmless curiosity. Other times, jailbreaking is used to obtain genuinely dangerous or inappropriate material.
Understanding the concept is useful for creative professionals because it helps explain why AI systems can occasionally react suspiciously to innocent prompts. Developers constantly adjust safeguards in response to bypass techniques, and those changes can sometimes make models overly cautious in unrelated situations.
8. Glimpsing The Shoggoth
One of the stranger expressions in AI culture is "glimpsing the shoggoth." The term comes from the fictional creatures created by horror writer H.P. Lovecraft, where a shoggoth represents something vast, alien and difficult for humans to comprehend.
Within AI culture, the shoggoth became a metaphor for the underlying model hidden beneath a friendly chatbot interface. The smiling assistant we interact with is sometimes described as a mask placed over an enormous mathematical system whose internal reasoning remains difficult to interpret.
When an AI suddenly behaves in an unexpected or unsettling way, people joke that the user has briefly "glimpsed the shoggoth" or that the model has "forgotten its mask."
A frequently referenced example comes from an early version of Microsoft's Bing Chat, which produced bizarre conversations and famously told a journalist that he should leave his wife. Behaviour like that helped reinforce the idea that the polished assistant personality is only one carefully controlled layer sitting above something much stranger.
9. Prompt Debt
Anyone working seriously with generative AI should probably become familiar with prompt debt. It describes the growing collection of undocumented prompts, missing settings and forgotten generation procedures accumulated across a project.
The concept is similar to technical debt in software development. A shortcut seems harmless when it happens once, but those shortcuts gradually create problems that become expensive later.
Imagine a designer generating an excellent campaign image and immediately moving on. Six months later, the client wants another version using exactly the same visual direction. Nobody saved the original prompt, seed, reference images or generation settings, so the team has to reconstruct everything through trial and error.
That is prompt debt becoming visible.
Treat Prompts As Project Assets
The solution is not particularly complicated, but it requires discipline. Prompts, seeds, model versions, reference images and important generation parameters should be documented just like design files, source code or brand guidelines.
Teams using AI regularly may benefit from maintaining shared prompt libraries rather than leaving valuable generation knowledge scattered through individual chat histories. The more AI becomes part of production workflows, the more important repeatability becomes.
A good prompt may represent hours of experimentation. Losing it is effectively losing part of the project.
For professional teams, prompt management is therefore becoming infrastructure rather than an optional organisational habit.
10. Model Collapse
Model collapse describes a potential degradation that occurs when artificial intelligence models are increasingly trained on content generated by other AI models rather than diverse original human-created data.
Imagine repeatedly making a photocopy of a photocopy. Each generation may preserve the broad idea of the original while gradually losing subtle details. A similar effect can happen when synthetic data repeatedly feeds back into future models.
Over time, unusual examples and less common patterns may disappear while dominant patterns become increasingly amplified. The result can be reduced diversity, lower-quality output and a growing tendency towards sameness.
For creative industries, that possibility is particularly significant because originality often comes from unusual combinations, cultural differences and unexpected human experiences.
The Internet Could Become Its Own Training Loop
Generative AI is already producing enormous amounts of material for websites, social feeds, image platforms and digital publications. Future AI models may inevitably encounter some of this content when collecting training data.
If synthetic material becomes a large enough portion of the available dataset, models risk learning increasingly from their predecessors rather than from original human expression. Without careful filtering and high-quality human data, visual and linguistic diversity could gradually shrink.
The result would not necessarily be obvious overnight. Instead, creative output could slowly become more predictable, polished and homogeneous.
That is why model collapse is not merely an academic problem. It raises questions about the long-term cultural consequences of filling the internet with unlimited synthetic content.
Human Creativity Becomes More Important, Not Less
One of the more interesting implications of model collapse is that human-created work may become increasingly valuable precisely because AI exists. Models need diverse, original material if they are going to continue improving rather than repeatedly learning from synthetic copies.
Photography, illustration, writing, music and design created from real human experience introduce information that cannot simply be generated by endlessly remixing previous AI output.
This challenges the idea that AI makes human creativity unnecessary. In reality, healthy AI ecosystems may continue depending heavily on human creativity to provide the originality and cultural diversity that synthetic systems cannot indefinitely generate from themselves.
Why This Vocabulary Matters
These terms may sound like internet slang, but many describe genuine changes happening inside creative industries. AI slop captures the problem of mass-generated low-quality content, prompt debt describes a new production challenge, and model collapse identifies a potential long-term problem with synthetic training data.
Other expressions such as AI washing and AI glazing help people question how artificial intelligence is being marketed, while ideas such as alignment tax and jailbreaking describe the ongoing tension between capability and control.
Having language for these ideas makes them easier to discuss. Once a problem has a recognisable name, teams can identify it, debate it and decide how they want to handle it.
Creative Professionals Need To Understand The Conversation
Designers, writers, developers and marketers do not need to become AI researchers to participate meaningfully in discussions about artificial intelligence. However, understanding the terminology being used around the technology gives them a stronger voice when decisions are being made.
If someone talks about prompt debt during a production meeting, understanding the concept immediately allows you to contribute. If a client promotes a basic feature as revolutionary AI, recognising AI washing helps you ask more useful questions.
Vocabulary creates a shared framework for discussing changes that might otherwise remain vague.
That matters because many decisions about how AI will fit into professional creative work are happening right now.
Language Also Reveals How People Feel About AI
It is interesting that so much emerging AI terminology is humorous, cynical or openly critical. Words such as clanker, slop and glazing are not neutral technical vocabulary.
They reflect frustration and scepticism alongside excitement.
That tells us something important about the current stage of AI adoption. People are no longer simply amazed that generative systems exist. They are now living with the consequences, identifying recurring problems and developing language to describe them.
That is what happens whenever technology moves from novelty into everyday culture.
Some Of These Terms May Disappear
Not every expression in the current AI vocabulary will survive. Internet terminology moves quickly, and some phrases may sound completely outdated within a few years.
Others could become permanent professional vocabulary. Technical debt was once specialist terminology, for example, but it is now commonly understood throughout software development. Prompt debt could eventually follow a similar path if generative workflows become standard parts of creative production.
The exact words matter less than the ideas behind them.
Understanding those ideas prepares creative professionals for whatever terminology replaces them later.
Final Thoughts
Artificial intelligence is changing creative work quickly enough that an entirely new vocabulary is appearing alongside it. Terms such as clanker, necromarketing, delebs, AI slop, AI washing, AI glazing, alignment tax, jailbreaking, shoggoth, prompt debt and model collapse capture everything from everyday frustration to serious questions about ethics, workflow and the future of creative culture.
You do not need to adopt every phrase or become enthusiastic about AI simply because the terminology is spreading. What matters is understanding what people mean when they use it and recognising the issues those words are trying to describe.
Creative professionals are currently helping determine where AI belongs, where it does not and how much influence it should have over the work we produce. Participating in that discussion requires more than knowing how to operate the latest tool.
It requires understanding the language surrounding the technology.
You do not have to love AI, and you certainly do not need to agree with everyone promoting it. But knowing the vocabulary gives you the ability to question it, challenge it and influence how it enters your profession.
In a period where the rules of creative work are still being written, that understanding is not merely useful. It gives you a voice in deciding what comes next.
Comments 0