Slopaganda | a new word for AI spin

What do ‘slithy’ and ‘mimsy’ have in common? Aside from sounding like pet names for English aristocrats, they’re both portmanteaus – words formed by blending two or more words that share similar sounds and meanings. Lewis Carroll popularised its literary use in Through the Looking-Glass (1871), where ‘slithy’ combined ‘lithe’ and ‘slimy’, and ‘mimsy’ fused ‘miserable’ and ‘flimsy.’ 

It’s a common way to describe new things or events, and love or hate it, we do it regularly. Personal favourites include ‘Volfefe,’ ‘Bennifer,’ ‘brunch,’ and ‘tomacco’ (clearly, some are more successful than others.)

At CMT this week, we had a good chortle about a new portmanteau describing a growing trend – ‘slopaganda’, a blend of 'AI slop' and ‘propaganda’ coined in a paper by Michał Klincewicz, Mark Alfano and Amir Ebrahimi Fard. As with any good portmanteau, it does what it says on the tin: it refers to unwanted AI-generated content aimed at manipulating beliefs for political ends.

Like traditional propaganda, slopaganda is designed to overwhelm the information environment with plausible-sounding material that distracts, misinforms or influences ideologies. The difference is, it’s distributed instantly and at an unprecedented scale using social bots and algorithmic mechanisms.

Its other key distinguishing feature is that slopaganda can be personalised in a way traditional propaganda cannot. With a TV programme or leaflet, it’s the same information for everyone. With slopaganda, messages can be tailored around a user’s characteristics.

This taps into our tendency to seek out repetition and confirmation bias, making it easier for slopaganda to flood newsfeeds and overshadow credible reporting. With advances in AI technology increasingly accessible, slopaganda can be created by anyone, anywhere, anytime. It’s no secret that terrorist organisations like ISIS are already using the technology for interactive recruitment, alongside groups looking to influence election outcomes.

Fighting slopaganda, or AI slop in general, may feel like a losing battle, but Klincewicz et al suggest that its impact could be curbed with a multi-layered strategy. This includes tactics like ‘prebunking’ (warning people in advance about misleading information) and introducing improved content moderation tools for AI detection. They even float the idea of a global wealth tax to limit the influence of corporations and oligarchs who use their capital to interfere with politics.

Despite the fun new name, our growing awareness of slopaganda – with its slithy knack for speed and mimsy half-truths – is a good thing. It reminds us that vigilance, media literacy and a commitment to journalistic rigour are more essential than ever.

Relevant links 

Portmanteau word, Definition, Origin, & Examples. (2025, May 14). Encyclopedia Britannica. https://www.britannica.com/topic/portmanteau-word‌

Perlman, M. (2016, April 18). ‘Chortle,’ and other words invented by Lewis Carroll. Columbia Journalism Review. https://www.cjr.org/language_corner/chortle.php

Willison, S. (2024, May 8). Slop is the new name for unwanted AI-generated content. Simon Willison’s Weblog. https://simonwillison.net/2024/May/8/slop/  

Klincewicz, M., Alfano, M., & Fard, A. (2025). Slopaganda: the interaction between propaganda and generative AI. Filosofiska Notiser, Årgång, 12, 135-162. https://www.filosofiskanotiser.com/KlincewiczAlfanoFard.pdf

Nelu, C. (2024, June 10). Exploitation of Generative AI by Terrorist Groups. International Centre for Counter-Terrorism - ICCT. https://icct.nl/publication/exploitation-generative-ai-terrorist-groups 

Stockwell, S., Hughes, M., Swatton, P., Zhang, A., & Hall, J. (2024, November). AI-enabled influence operations: safeguarding future elections. Centre for Emerging Technology and Security. https://cetas.turing.ac.uk/sites/default/files/2024-11/cetas_research_report_-_ai-enabled_influence_operations_-_safeguarding_future_elections.pdf