The ‘I’ in AI

• Artificial Intelligence (AI) seems to be everywhere. Almost any action on an online computer or smartphone seems to result in a chatbot, suggested edit or an offer to research your topic. Our interactions with the dentist, doctor, bank or plumber are increasingly mediated through AI.

Yet behind all the discussion about what AI can do lies a more important question: can there ever be an ‘I’ in AI?

There is a lot to like about AI. It is like an extraordinarily fast research assistant, capable of gathering, organising and presenting vast amounts of information in seconds. It saves huge amounts of time. It opens new possibilities for people functioning outside their language of comfort in its ability to instantly translate between languages. On and on the advantages go.

However, there are also downsides to AI. Communities divide over the way in which AI data centres divert scarce resources like water and electricity. Intellectual theft and deception abound when AI outputs are claimed by a human person. Development of personal creativity and critical thinking can be hindered by laziness and a quick resort to AI-based research. Some predict that as AI feeds on its own increasing outputs it may become increasingly self-referential. AI can open dark possibilities of privacy breaches, unwanted surveillance and cybercrime when in the wrong hands – let alone the ways in which it feeds the porn industry.

AI also raises questions about the nature of people.

In a recent test, a standard ethical dilemma was fed into AI. It did a great job of analysing the relevant issues and outlining how various approaches to ethics could inform different responses to the dilemma. The question was then asked: ‘What should I do?’. The key word there is ‘should’, for it asked AI to make an ethical judgement. The answer was interesting: ‘I cannot tell you what you should do as I am not a person.’ The key word there was ‘I’. AI here presents itself as an “I”, even though it possesses no personal consciousness and says that it cannot make a moral judgement. That word ‘I’ risks blurring the distinction between a machine and a person.

Should we mistake AI’s use of the word “I” for genuine personhood?

Christians understand people to be more than physical bodies. We have bodies, but we also have the inward aspects of mind, heart and soul. Some current understandings of the human mind see it as an essentially and purely material organ – a bunch of electrical connections and chemical reactions. If that is so, we can see why AI is often compared to the workings of the mind. (It’s worth noting that robots equate to a ‘bodily’ extension of AI.)

But what of heart and soul – those inward aspects that relate to love, fear, worship, appreciation of beauty and such like? These are all important parts of human meaning and joy – they make us ‘us’. They also distinguish us from animals and machines. How can a purely materialistic understanding of humanity fully account for love, beauty, worship, moral responsibility and hope?

It’s a grim and inadequate view of people to reduce us to a purely machine-like nature. A Christian view of people makes far better sense. According to Genesis, humanity alone is created in God’s image and entrusted with responsibility over creation. We believe that this gives every human being a dignity and calling that no machine can possess. AI is an extraordinary tool, but it is not a person. We can value its usefulness without forgetting the unique dignity of the human beings who created it – and who alone bear the image of God.

There may be an ‘I’ in the term ‘AI’, but the true “I” belongs only to people made in the image of God.