A Good Year for the Machines
The assistant on my phone tells me that 2016 was its year – “The Year of AI!” Many articles with such titles have been suggested to me recently. The funny thing is that algorithms probably wrote some of these articles. Coincidence? Or are these artificial intelligences conspiring to get more media coverage?
A bad joke, I admit, but you just never know these days – you can be interacting with an AI and not even know it: a while back I read that Deepmind’s Go-playing algorithm, ‘AlphaGo’, had been playing online under an anonymous pseudonym, ‘Master’. The mysterious new player gathered media attention around the world by maintaining an unbeaten record against some of the best players in the game.
Machines are not just writing banal news pieces about the weather or sports results, they are writing novels and screenplays. An AI screenwriter called Benjamin recently entered its first short film, Sunspring, into the Sci-Fi London Film Festival last year. Benjamin has learnt to ‘write’ by ‘reading’ countless scripts from science fiction films and television shows. Its script is terrible, sure, but it is only the beginning – people thought that AlphaGo wasn’t very good a few years ago, and now, no one will probably ever beat it again.
AI is rendering three-dimensional figures for film: for example, companies use algorithms in order to help render figures in the background during scenes with large numbers of people. Deepmind has won a court case in the UK that grants them access to millions of NHS patients’ records – they aim to create machines that can diagnose various forms of illness. For example, one is already able to accurately diagnose forms of eye-cancer as well as their human counterparts. A poker-playing AI, Liberatus, just won a tournament against human opponents. Its victory in a game that involves ‘incomplete information’ is considered very important and shows that machines are becoming capable of understanding complex game theory.
Public awareness of AI is also increasing due to a number of recent television series, such as Westworld, Humans and Black Mirror. Each has explored and helped to popularize some of the sticky ethical questions raised by the prospect of artificial intelligence. For example, what rights should belong to beings that think and feel in the same way as we do? Westworld and Black Mirror, in particular, explore some of the incredibly dark possibilities that arise when such beings are not given any rights.
Then, of course, there is the broader question of what it means to be human. When the artificial thinks and feels and looks like a human, what does it mean to be real? The answer, as far as Westworld is concerned, is freedom of will, about listening to our ‘inner voice’. The series suggests that we are the stories that we tell ourselves, and the ability to tell the kind of story that we want. However, it goes deeper than this almost prosaic existentialist maxim: Ford points out that humans, like the hosts, are also ‘stuck in loops’ – we may think that we are free, but in reality, we are at the mercy of forces beyond our control and even understanding.
Watching Westworld, I found myself wondering about the type of world that exists outside of the park, a world which we never really see: what kind of a society allows such a thing? What has become of humankind that they allow self-conscious entities that can feel pleasure and pain to be tortured, raped and killed?
The problem is that we are not so far behind: we already treat the other species on this planet appallingly, and while (the majority of us) may not torture animals, we are certainly complicit in their suffering and death, and this mostly serves to provide us with pleasure (given that we don’t have to eat meat).
I don’t think we should even attempt to create AI indistinguishable from humans: at worst, we will fail, and at best, we will end up creating entities without rights – we don’t have a good enough track record to think that anything else would be the case.