Reposted by joeposaurus
(2/7) LLMs are a dark pattern w/ no inherent affordances or scope whose utility depends on user affect. Unlike other tools LLMs give no feedback when you're using them deleteriously. Also, you have to trust that no designer takes advantage of user affect to shape user behavior.
misaligned.markets
LLM “intelligence” is a dark pattern
LLMs leverage scaffolding and user psychology to appear as slabs of raw intelligence in service of a costly illusion.
![Another related aspect — described by Valentina Ponomareva (1998) — is that:
it is impossible to create an absolutely reliable automatic system, and sooner or later people face the necessity to act after equipment fails. [...] If the cosmonaut loses such skills because of [their] passive role [due to being typically limited to monitoring and observation only], the probability of [their] choosing and carrying out the right procedure in an emergency would be small. This contradiction is inherent in automatic control systems.
For more on this angle of AI see:
Guest, O. (2025). What Does 'Human-Centred AI' Mean?. arXiv. https://doi.org/10.48550/arXiv.2507.19960](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:vsyj5jtayxcq5dw5mtja4xtb/bafkreigj5dinoh57bo6zaslraebhe67rifpibyo4u7r6tmzll5xnyelo7e)
