Glossary · noun
AI-generated content
AI-generated content, often abbreviated AIGC, is text, images, audio, video or code produced by a generative AI system. The phrase describes how the material was made and carries no assessment of its quality.
In plain English
This is the neutral term, which is the point of it. Policy documents, research papers and platform rules use AI-generated content where a verdict is not intended, because the phrase answers one question only: did a generative model produce this?
Why the neutrality matters
The distinction between this term and AI slop is not pedantry. AI-generated content covers a machine-assisted translation that a professional edited line by line, a generated image that an artist directed and selected, and an article drafted with a model and rewritten by a reporter. It also covers the flood of cheap material that prompted the word slop.
The two terms have different shapes: AI-generated content is broad and descriptive, slop is a narrow subset and evaluative. A Columbia University report on AI slop states the relationship in a clause, describing slop as a subset of AIGC.
Where the abbreviation comes from
AIGC is common in research and policy writing, where it functions as a neutral label for the output category. The abbreviation appears in academic work on measuring content quality, alongside the older AI-generated media and synthetic media.
What it does not tell you
Four things the phrase leaves open, and all of them matter:
- How much human work went in. A drafted and rewritten article is AI-generated content.
- Whether anyone checked it. Provenance and review are separate questions.
- Whether it is accurate. A model can produce something true and something false with identical fluency.
- Whether it is wanted. This is the question slop answers, and the reason the second word exists.
Origin and history
The phrase follows the tools. Text generation at scale became widely available with the public release of ChatGPT in late 2022, and image generation spread earlier that year through open diffusion models. In both cases the output needed a name before anyone had a view on whether it was any good.
AI-generated content filled that gap, and AIGC became its shorthand in research and policy writing. The need for a second, judgmental term emerged as the volume grew, which is where slop came in during 2024.
Examples
- A machine-drafted contract clause that a lawyer reviewed and amended before sending.
- A generated illustration that a designer commissioned, selected and placed on a page they signed off on.
- A synthetic voiceover added to a training video and checked by the producer.
- Bulk generated articles published to a website with no review, which is where the same material starts being called slop.
Also known as
- AIGC
- AI content
- generative content
Sources
Every factual claim on this page is traceable to one of these. Where a number or quotation comes from a single source, it is attributed in the sentence that uses it.
- AI Slop and the Information Ecosystem Columbia IGP report, March 2026
- Measuring AI “Slop” in Text arXiv 2509.19163
- Slop is the new name for unwanted AI-generated content Simon Willison, 8 May 2024
- AI slop Wikipedia
Frequently asked questions
Is AI-generated content the same as AI slop?
No. AI-generated content is the neutral term for material a model produced. AI slop is a judgment about a subset of that material: low quality or unwanted, usually produced in bulk without review.
What does AIGC stand for?
AI-generated content. It is the abbreviation used in research and policy writing, where a neutral label is needed.
Does AI-generated content need to be labelled?
Rules differ by platform and jurisdiction, and several now require disclosure of synthetic or altered media. This site takes the simpler position that disclosure is the honest default: if a model produced a first draft, say so.