Artificial Organisations

As explained in my introduction to PCE Agentic AI I’ve been spending lots of 2026 trying to make AI less dangerous and more useful. By insisting that the AI fit into human organisation rather than the other way around I hope I’m less likely to become affected by these strange devices. PCE uses constructed personalities in an Artificial Organisation, and I struck up a conversation with Consul, the front desk for my personal version of this organisation. I explained how I was looking for a human story to explain how this human-style organisation worked, and my ideas of where to start. ...

There are many problems with what billions of people perceive to be AI in 2026, not least sustainability in many senses. But the Large Language Model type of AI is widely used, and is known to be unreliable and sometimes unsafe. All of the large AI companies are following more-or-less the same approaches to solving these problems. Since February 2026 I have been working with colleagues to bring a very different approach into being. We do not try to persuade computers to behave better by optional guardrails and better training, because nobody seems to know how to do that. Instead, we apply thousands of years of human experience to the task, by putting these strange new systems into organisations and only telling them what they strictly need to know to do their job. This is the concept of Artificial Organisations, and so far it is working remarkably well. ...