What is the EU doing on AI regulation that could affect firms using AI in fiduciary areas like tax, law, and finance?

In Europe, there is heavy emphasis on defining what counts as a 'high-risk' AI system. Activities involving elements of social scoring are classified as super high risk and can trigger fines. This month (as discussed), the official EU AI Act was being released, with a compliance deadline referenced as June 2026, requiring every company to comply with certain legislation and get vetted by the AI board in their country. One firm described being assessed as 'low risk' because they were not automatically sending data to a government authority - they kept a human in the loop. If they later built a connector to a European entity with automatic submission capability, that would reclassify them as high risk, exposing them to potential fines. The overall take is that regulators are trying to protect people, but this comes at a trade-off with stifling innovation.