It's our mistakes that make us human

What we learn distinguishes us from tech.

Silvianne Aspray

Silvianne Aspray is a theologian and postdoctoral fellow at the University of Cambridge.

A man staring at a laptop grimmaces and holds his hands to his head.

Francisco De Legarreta C. on Unsplash.

The distinction between technology and human beings has become blurry: AI seems to be able to listen, answer our questions, even respond to our feelings. It becomes increasingly easy to confuse machines with humans. In this situation, it is increasingly important to ask: What makes us human, in distinction from machines? There are many answers to this question, but for now I would like to focus on just one aspect of what I think is distinctively human: As human beings, we live and learn in time.

To be human means to be intrinsically temporal. We live in time and are oriented towards a future good. We are learning animals, and our learning is bound up with the taking of time. When we learn to know or to do something, we necessarily make mistakes, and we take practice. But keeping in view something we desire – a future good – we keep going.

Let’s take the example of language. We acquire language in community over time. Toddlers make all sorts of hilarious mistakes when they first try to talk, and it takes them a long time even to get single words right, let alone to try and form sentences. But they keep trying, and they eventually learn. The same goes with love: Knowing how to love our family or our neighbours near and far is not something we are good at instantly. It is not the sort of learning where you absorb a piece of information and then you ‘get’ it. No, we learn it over time, we imitate others, we practice and even when we have learned, in the abstract, what it is to be loving, we keep getting it wrong.

This, too, is part of what it means to be human: to make mistakes. Not the sort of mistakes machines make, when they classify some information wrongly, for instance, but the very human mistake of falling short of your own ideal. Of striving towards something you desire – happiness, in the broadest of terms – and yet falling short, in your actions, of that very goal. But there’s another very human thing right here: Human beings can also change. They – we – can have a change of heart, be transformed, and at some point in time, actually start to do the right thing – even against all the odds. Statistics of past behaviours, do not always correctly predict future outcomes. Part of being human means that we can be transformed.

Transformation sometimes comes suddenly, when an overwhelming, awe-inspiring experience changes somebody’s life as by a bolt of lightning. Much more commonly, though, such transformation takes time. Through taking up small practices, we can form new habits, gradually acquire virtue, and do the right thing more often than not. This is so human: We are anything but perfect. As Christians would say: We have a tendency to entangle ourselves in the mess of sin and guilt. But we also bear the image of the Holy One who made us, and by the grace and favour of that One, we are not forever stuck in the mess. We are redeemed: are given the strength to keep trying, despite the mistakes we make, and given the grace to acquire virtue and become better people over time. All of this to say that being human means to live in time, and to learn in time.

So, this is a real difference between human beings and machines: Human beings can, and do strive toward a future good.

Now compare this to the most complex of machines. We say that AI is able to “learn”. But what does it mean to learn, for AI? Machine learning is usually categorized into supervised learning, unsupervised and self-supervised learning. Supervised learning means that a model is trained for a specific task based on correctly labelled data. For instance, if a model is to predict whether a mammogram image contains a cancerous tumour, it is given many example images which are correctly classed as ‘contains cancer’ or ‘does not contain cancer’. That way, it is “taught” to recognise cancer in unlabelled mammograms. Unsupervised learning is different. Here, the system looks for patterns in the dataset it is given. It clusters and groups data without relying on predefined labels. Self-supervised learning uses both methods: Here, the system uses parts of the data itself as a kind of label – such as, for instance, predicting the upper half of an image from its lower half, or the next word in a given text. This is the predominant paradigm for how contemporary large-scale AI models “learn”.

In each case, AI’s learning is necessarily based on data sets. Learning happens with reference to pre-given data, and in that sense with reference to the past. It may look like such models can consider the future, and have future goals, but only insofar as they have picked up patterns in past data, which they use to predict future patterns – as if the future was nothing but a repetition of the past.

So this is a real difference between human beings and machines: Human beings can, and do strive toward a future good. Machines, by contrast, are always oriented towards the past of the data that was fed to them. Human beings are intrinsically temporal beings, whereas machines are defined by temporality only in a very limited sense: it takes time to upload data, and for the data to be processed, for instance. Time, for machines, is nothing but an extension of the past, whereas for human beings, it is an invitation to and the possibility for being transformed for the sake of a future good. We, human beings, are intrinsically temporal, living in time towards a future good – which machines do not.

In the face of new technologies we need a sharpened sense for the strange and awe-inspiring species that is the human race, and cultivate a new sense of wonder about humanity itself.

Article

Culture

14 March 2024

5 min read

What AI needs to learn about dying and why it will save it

Those programming truthfulness can learn a lot from mortality.

Andrew Steane

Andrew Steane has been Professor of Physics at the University of Oxford since 2002, He is the author of Faithful to Science: The Role of Science in Religion.

An angel of death lays a hand of a humanioid robot that has died amid a data centre

A digital momento mori.

Nick Jones/midjourney.ai

Google got itself into some unusual hot water in recently when its Gemini generative AI software started putting out images that were not just implausible but downright unethical. The CEO Sundar Pichai has taken the situation in hand and I am sure it will improve. But before this episode it was already clear that currently available chat-bots, while impressive, are capable of generating misleading or fantastical responses and in fact they do this a lot. How to manage this?

Let’s use the initials ‘AI’ for artificial intelligence, leaving it open whether or not the term is entirely appropriate for the transformer and large language model (LLM) methods currently available. The problem is that the LLM approach causes chat-bots to generate both reasonable and well-supported statements and images, and also unsupported and fantastical (delusory and factually incorrect) statements and images, and this is done without signalling to the human user any guidance in telling which is which. The LLMs, as developed to date, have not been programmed in such a way as to pay attention to this issue. They are subject to the age-old problem of computer programming: garbage in, garbage out.

If, as a society, we advocate for greater attention to truthfulness in the outputs of AI, then software companies and programmers will try to bring it about. It might involve, for example, greater investment in electronic authentication methods. An image or document will have to have, embedded in its digital code, extra information serving to authenticate it by some agreed and hard-to-forge method. In the 2002 science fiction film Minority Report an example of this was included: the name of a person accused of a ‘pre-crime’ (in the terminology of the film) is inscribed on a wooden ball, so as to use the unique cellular structure of a given piece of hardwood as a form of data substrate that is near impossible to duplicate.

The questions we face with AI thus come close to some of those we face when dealing with one another as humans.

It is clear that a major issue in the future use of AI by humans will be the issue of trust and reasonable belief. On what basis will we be able to trust what AI asserts? If we are unable to check the reasoning process in a result claimed to be rational, how will be able to tell that it was in fact well-reasoned? If we only have an AI-generated output as evidence of something having happened in the past, how will we know whether it is factually correct?

Among the strategies that suggest themselves is the use of several independent AIs. If they are indeed independent and all propose the same answer to some matter of reasoning or of fact, then there is a prima facie case for increasing our degree of trust in the output. This will give rise to the meta-question: how can we tell that a given set of AIs are in fact independent? Perhaps they all were trained on a common faulty data set. Or perhaps they were able to communicate with each other and thus influence each other.

The questions we face with AI thus come close to some of those we face when dealing with one another as humans. We know humans in general are capable of both ignorance and deliberate deception. We manage this by building up degrees of trust based on whether or not people show behaviours that suggest they are trustworthy. This also involves the ability to recognize unique individuals over time, so that a case for trustworthiness can be built up over a sequence of observations. We also need to get a sense of one another's character in more general ways, so that we can tell if someone is showing a change in behaviour that might signal a change in their degree of trustworthiness.

In order to earn our trust, an AI too will have to be able to suffer and, perhaps, to die.

Issues of trust and of reasonable belief are very much grist to the mill of theology. The existing theological literature may have much that can be drawn upon to help us in this area. An item which strikes me as particularly noteworthy is the connection between suffering and loss and earning of trust, and the relation to mortality. In brief, a person you can trust is one who has ventured something of themselves on their pronouncements, such that they have something to lose if they prove to be untrustworthy. In a similar vein, a message which is costly to the messenger may be more valuable than a message which costs the messenger nothing. They have already staked something on their message. This implies they are working all the harder to exert their influence on you, for good or ill. (You will need to know them in other ways in order to determine which of good or ill is their intention.)

Mortality brings this issue of cost to a point of considerable sharpness. A person willing to die on behalf of what they claim certainly invests a lot in their contribution. They earn attention. It is not a guarantee of rationality or factual correctness, but it is a demonstration of commitment to a message. It signals a sense of importance attached to whatever has demanded this ultimate cost. Death becomes a form of bearing witness.

A thought-provoking implication of the above is that in order to earn our trust, an AI too will have to be able to suffer and, perhaps, to die.

In the case of human life, even if making a specific claim does not itself lead directly to one's own death, the very fact that we die lends added weight to all the choices we make and all the actions we take. For, together, they are our message and our contribution to the world, and they cannot be endlessly taken back and replaced. Death will curtail our opportunity to add anything else or qualify what we said before. The things we said and did show what we cared about whether we intended them to or not. This effect of death on the weightiness of our messages to one another might be called the weight of mortality.

In order for this kind of weight to become attached to the claims an AI may make, the coming death has to be clearly seen and understood beforehand by the AI, and the timescale must not be so long that the AI’s death is merely some nebulous idea in the far future. Also, although there may be some hope of new life beyond death it must not be a sure thing, or it must be such that it would be compromised if the AI were to knowingly lie, or fail to make an effort to be truthful. Only thus can the pronouncements of an AI earn the weight of mortality.

For as long as AI is not imbued with mortality and the ability to understand the implications of its own death, it will remain a useful tool as opposed to a valued partner. The AI you can trust is the AI reconciled to its own mortality.