Hook
More breakout videos from this creator.
In 2016, Dave and I analyzed over nine hundred occupations in a job study. We were trying to figure out what skills kids should build to stay competitive against automation and we found four clusters of work that machines couldn't touch. The common factor amongst them was unpredictability. We had messy environments, physical spaces, spaces where the data runs out and judgement begins. We weren't wrong and that's done really well over the years but we now want to update our advice. The automation we were modelling in 2016 was very much just pattern matching, process forms and recognise images - that kind of automation that chipped away at routine work and left analysis, strategy, language and creativity alone. Now generative AI handles plenty of unpredictable situations. It improves and adapts and generates responses to prompts it's never seen. But what it can't do is five different kinds of difficulty together at once in a physical place with real consequences when no one agrees what success looks like. Unpredictability was a symptom but the underlying condition is actually complexity. A specific kind of complexity that resists handed to AI and reassembled. Now that's what this series is about. So I went back to the data. Another almost 900 occupations in the US and asked a different question. What makes work irreducibly complex? Five dimensions came out of the data and that's what this series is about. I'm Helen from Artificiality. Follow me for more on staying human with AI.