Stack Overflow, the developer Q&A site best known for its commitment to human expertise and high-quality answers, confirmed this week that over 80% of its users are now using AI tools like ChatGPT to answer coding questions — while simultaneously insisting they have never once trusted anything AI-generated.

A recent poll conducted by Stack Overflow’s new Algorithmic Trust Deficit Division found that only 29% of users believe AI tools are even remotely reliable. “I would sooner let a squirrel write my API documentation than trust AI,” declared longtime user ‘NullPointerException’, before quietly copy-pasting a ChatGPT answer into production code. “I come to Stack Overflow for human wisdom, then I ask GPT-4 because it’s faster,” he added.

CEO Prashanth Chandrasekar downplayed concerns about AI-driven hypocrisy, stating, “Stack Overflow remains the trusted home for developers who want to ask humans what to do after AI has already made all their decisions for them.” He announced future plans to introduce an AI-powered Stack Overflow feature called ‘BlameBot’, which will automatically accept all liability for hallucinated answers, freeing up developers to focus on more important tasks, like clicking ‘Accept Solution.’

Veteran moderator Susan McSemicolon summarized the mood: “AI is like my ex—unfaithful, unreliable, and yet here I am, texting it at 2 a.m. to fix my CSS.”

In related news, Stack Overflow’s annual conference will feature workshops such as “How To Hate AI While Secretly Using It For Literally Everything” and “Copy-Paste Therapy: Healing From Code You Didn’t Write.”