In May 2026, the European Commission published its draft guidance on high-risk AI systems and pushed the compliance date from August 2026 to December 2027. The same week, Verizon's annual Data Breach Investigations Report showed that 45 percent of employees now use AI on corporate devices, mostly through personal accounts. And the Pay Transparency deadline is just days away.
In this episode, Anna Carlsson breaks down what the AI Act update actually says, why the pushed deadline is more risk than relief for most Nordic organisations, and why Pay Transparency and the AI Act aren't two separate compliance projects. They're the same data challenge in two forms. The episode also covers what happens when you build your own AI solutions - custom GPTs, Copilot agents, Claude-based assistants - and why fine-tuning a foundation model for HR decisions can turn you into a "provider" under the AI Act. Plus four practical steps to take now, instead of waiting until 2027.
