Description

This field note captures Abi’s reaction to Facebook trialling AI-generated prompts and suggestions in the feed, with 20 examples that range from mildly confused to existentially cursed.

The carousel looks at what happens when AI is layered into a social experience without enough understanding of context, humour, tone, intent or emotional nuance. The examples include prompts that misread jokes, flatten absurdity, offer unsolicited explanations, and turn obviously funny or chaotic posts into bizarre recommendation prompts. As Abi puts it early on, it feels like Facebook invited “a party pooper” into the feed: one that does not quite understand humour, lacks context, loves unsolicited advice, and is funny for the wrong reasons.

Under the comedy, there is a serious UX point: just because AI can be added to an interface does not mean it improves the experience. Poorly judged AI can interrupt the social rhythm of a product, misunderstand user intent, reduce delight, and make the whole system feel less human.

Topics

• AI-generated prompts in social feeds

• Context failure and humour misinterpretation

• Unsolicited AI assistance

• Poor timing and weak relevance

• When automation damages the experience

• The difference between useful AI and performative AI

• Why “just because you can” is not a product strategy

• Real-world examples of AI creating friction, confusion and accidental comedy

Best for

AI, Facebook, UX, product design, social media, context, humour, automation, interface design, field notes, AI UX

Background

This piece sits alongside Abi’s wider work on UX, AI and trust because it shows the gap between technical capability and meaningful usefulness. The issue is not whether AI can generate prompts. The issue is whether those prompts make sense in the user’s actual context.

It also connects to the Corpus view that machine interpretation needs to be judged against human reality. When AI misreads humour, tone or intent, the interface stops feeling helpful and starts exposing the system’s lack of understanding. That is not enhancement. That is a context failure with a shiny badge.