2 September

Everything is in the framing. If we see AI as some magical black box that contains 'intelligence', or just statistical tricks, everything we think and say about it will be framed by that assumption. Every time it gets something 'wrong' it will be seen as unreliable, as a risk. Or if it seems too agreeable, it will be seen as 'hallucinating' the things we want to see. Neural networks learn. There is no black box. Learning *is* making mistakes. It *is* emulating those we're trying to learn from, because that's how we learn. We need a reference point to test if we're doing it right. We need to stop talking about AI as some other arriving in our societies. It's the natural evolution of networked knowledge, creating neuronal nodes in those networks that allow them to think (yes think) and act within those networks.