I’ve got a bit more time now @Simon Ozanne, so I can put a bit more thought into a response. This whole post started as a result of frustration. We’ve been swimming backwards in our work due to increased surveillance requirements by our new one and two stars. They are way more curious than previous principles, which is great, but that curiosity is creating a level of fear/urgency/unrequited churn. This coincides with the release and heavy use of copilot, which is generating another way of reporting. The one thing we’re very aware of however, is to ensure AI doesn’t become another form of reporting without improvement. A lot of organisations already collect way more information and they can act on. We’re trying to make sure that AI doesn’t increase the problem by producing more themes, more alerts and more dashboards without increasing our capacity to respond.. We’re starting to realise that the real measure of AI in safety, is not how much it detects. The real measure will be how much we can use it to improve the work. If AI helps an organisation see more but improve nothing, it is not learning. We are working out very quickly that observation without obligation, is a very slippery slope. Are you seeing the same thing in your work @Hilbert Monsuur