fr just saw a post arguing that companies are failing at ai governance bc they rely on paperwork instead of engineering. you can't stop people from using unauthorized tools just by handing out a handbook that nobody ever opens. the only real way to curb shadow ai is to make the
approved workflows actually
easier than the alternatives. developers will always take the path of least resistance, sooo we need to build responsible guardrails directly into the pipeline. if the official setup is a hassle, people will just bypass it.
it's basically an ux problem, not a legal one . anyone else seeing teams struggle w/ this rn?
https://stackoverflow.blog/2026/08/24/responsible-ai-adoption-needs-developer-workflow-design/