2023 is the year artificial intelligence finally grabbed peoples’ attention and sparked a market boom, perhaps even a bubble, along with dire warnings about mass unemployment and even the end of civilisation. So how should non-experts think about it? Something super scary and attention-grabbing that needs regulating? Or something to sit back and enjoy – the start of a new golden age?
Let’s get the difficult stuff over with first.
The idea of AI has been around for almost 70 years, and that’s part of the problem. Writers and technologists envisioned a future of artificial general intelligence (AGI) – leading to conscious god-like machines with a choice between good or evil. We’ve already seen the movies.
But this foundation myth leaves us ill equipped to understand the world that is actually emerging. Before reaching for their box of Blade Runner and Terminator metaphors, anyone who writes about AI needs to think carefully about the agendas embedded in their choice of subject and language.
Start with the terminology. In sci-fi writer Ted Chiang’s words the original sin was the term AI itself, and the casual read-across of human terms like ‘learning’. What we currently have according to Chiang is really ‘applied statistics’ (which is actually a reference to the triumph of neural networks but we’ll discuss that in a moment).
Let’s also not forget the invisible hand of capitalism. AI is a hype machine. It keeps everyone obsessed with giants like Google, Meta and Microsoft – and newbies like OpenAI – that have the resources to invest, and also benefit from regulation that protects their dominance. Your attention helps channel investments into countless start-ups seeking to bring the world chatbot-powered teddy bears and so on. Follow the money.
And bear in mind that AI is already accessible like no other scientific development. It’s not like gravitational wave detectors or virus research that require special labs and teams of scientists sequestered from the public eye. Much of the software is open-source, available to anyone with a decent computer and some coding skills. The key scientists participate in podcasts and run their own Twitter accounts. Research is being generated at an exponential rate – some 5,000 AI papers a month published on open-source repository Arxiv, doubling every 24 months.
Faced with all that, how is a writer going to tame the fire hose of digital information and produce a book, printed on 600-year-old technology, an actual printing press?
Thankfully, a number of journalists and experts have already attempted it. Let’s go through them and learn what they say, and what is still left to be written.
