Reckless Intelligence is a way for me to explore the unlimited, unbridled intelligence of humanity, in all its glory and all its folly – and to explore what’s happening as we build artificial intelligence that mirrors that recklessness.
I’m not concerned that AI will replace a lot of workers, will make a lot of jobs redundant. Of course it will. Do I care? Yes, of course. But am I resigned to the fact? Yes, as we all should be.
What I’m concerned about is the existential threat of AI. What happens if malicious agents have better AI than altruistic ones, and the willingness to use it to get whatever they want? Because they do. What happens if our military, financial and logistical systems become dependent on AI models that we can’t fully trust? Because they will.
AI is a wonderful tool with potential to cure disease, to help humans thrive. To come up with strategies to mitigate and maybe even reverse climate change. To make sense of our complex financial systems in a way that spreads wealth around. In short, to help us solve the world’s most complicated problems.
But even the large language models that are so widely available for free to all – they’re not coded with a simple goal in mind. They’re complex neural networks. They may have been asked to solve a very simple problem; but they figure out how to do that all by themselves. We have no way of knowing how they’re doing what they’re doing, no way of unpicking the threads of this messy tapestry, and certainly no way of anticipating each and every way it could go stunningly, disastrously wrong.
So here I am, asking all of the questions I can think of about this reckless intelligence, and exploring the answers. I’ll look anywhere. AI research. Futures Studies. Science fiction. Wherever the imagination exists to figure out what’s happening with the most groundbreaking, paradigm-shifting, world-upturning technology we’ve created in centuries.
Strap in.
reckless intelligence 