Glossary · noun
AI spam
AI spam is unsolicited bulk content generated with AI tools, produced to reach a large audience, an inbox or a search ranking system, typically for commercial gain. It is where the older category of spam meets the newer economics of generative content.
In plain English
Spam has meant unwanted bulk email since the 1990s. AI spam keeps the unwantedness and the bulk and changes the production method: models make the material, which drops the marginal cost of each attempt close to zero.
AI spam versus AI slop
The two overlap and neither contains the other.
- Spam needs a target. Whether it aims at an inbox, a comment section or a search result, spam is trying to reach a specific audience and usually wants an action: a click, a payment, a credential.
- Slop needs a slot. A generated video posted for platform revenue may reach almost nobody and still be doing its job, because one success pays for hundreds of attempts.
That is why a generated marketing email full of invented claims counts as both, while a generated video about a cat playing chess counts only as slop.
Where the volume comes from
Three enablers, and all three are about cost:
- Production cost collapsed. A page, an image, a voiceover or a product listing can be made in seconds.
- Distribution is nearly free. Social platforms, marketplaces and open comment sections let anyone publish to anyone at no marginal cost.
- Something pays per impression. Advertising revenue, affiliate funnels and attention arbitrage turn volume into income.
Remove any one of those and the volume falls. That is why attempts to address AI spam by filtering content have a mixed record, and why writing on the topic keeps returning to platform incentives.
Origin and history
The word spam comes from a 1970 Monty Python sketch and was adopted for unwanted bulk messages in the early internet era, then for email specifically as it became a mass medium in the 1990s.
AI spam is a straightforward extension once generation became cheap. Research and reporting on AI slop repeatedly note the economic drive behind the volume: material produced to game recommendation and ranking systems, with revenue arriving through advertising, affiliate links or funnels.
A Columbia University report on AI slop and the information ecosystem describes networks of pages using generated imagery and text to generate engagement at scale, often routing people toward monetized spam, data-harvesting sites or low-quality commercial funnels. That description is AI spam by any definition.
Examples
- A generated blog post published only to rank for a commercial keyword.
- Fake product listings with synthetic photographs on a marketplace.
- Generated comments and replies from bot accounts building the appearance of interest.
- Machine-written outreach at scale with invented claims about the sender.
- Synthetic news-style pages whose only purpose is to carry advertising.
Also known as
- slop spam
- AI-generated spam
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
- AI slop Wikipedia
- Slop is the new name for unwanted AI-generated content Simon Willison, 8 May 2024
- AI Slop: Definitions and Normative Status MINT Lab
Frequently asked questions
Is AI spam illegal?
Some of it is. Fraud, deceptive advertising and bulk messaging rules already cover many of the practices involved, and several jurisdictions regulate unsolicited commercial messages directly. Generating content with a model is not itself unlawful.
How is AI spam different from ordinary spam?
The production method and the cost. Ordinary spam required a human to compose each message or template. AI spam can be generated per recipient, per keyword and per channel at a cost near zero.
Is AI spam the same as SEO spam?
Heavily overlapping. Content written to rank has existed for as long as search engines have, and models made producing it dramatically cheaper, which is why the two are now discussed together.