What ‘Artificial Intelligence’ can teach us about humanity

Lynne Bowyer and Deborah Stevens

The term ‘Artificial Intelligence’ (AI) was coined by the computer scientist John McCarthy in 1956. It is an area of computer science involved with the development of computer systems capable of performing tasks normally requiring aspects of human intellect. We are currently seeing exponential growth in AI, and this is due to a number of factors: the availability of faster hardware; the ability of computers to run more complex algorithms; the development of massive data sets; and the enormous monetary incentives involved with this technology.

Our lives are already permeated by a range of what has been called ‘Narrow AI’ applications. ‘Narrow AI’ is incorporated into many of the products and procedures that impact on our daily lives: automatic heat pumps; smart-phones; web-based searches; self-driving cars; complex assembly work in industrial processes; Facebook automatically labelling your friends in photos; Amazon and Netflix making personalised product and film recommendations. Already these ‘Narrow AI’ applications do calculative tasks faster than humans can, and are altering our world in significant ways. The increase in speed and capability of ‘Narrow AI’ has enabled increased flexibility within AI systems so that in many cases those systems are able to transfer what they ‘learn’ in one domain to another. For those working in the field, this is fuelling the idea that we can create what has been termed ‘Artificial General Intelligence’(AGI).

AGI technologies have been built that incorporate algorithms that mimic aspects and degrees of human cognitive function, including visual perception, speech recognition and means-ends decision-making. This field of research draws on experiments with ‘machine learning’ that utilise neural network technology. This technology creates simplified models of brain networks that can self-organise and solve problems. With such things as voice and image recognition, and faster computational power, it is envisaged that AGI will be as capable as any human across any ‘intellectual task’. This is said to include things like complex reasoning, thinking abstractly and learning from experience.

These claims and the impetus behind these developments give us pause for thought. They give us the opportunity to ask the questions: What is human intelligence? How is human intelligence developed? Can human intellect ever be simulated by a computer programme?

We argue that human intellect cannot be simulated by a computer programme, however complex that programme may become. This is because human intelligence is more than being able to access and process information in an abstract way and calculate means-ends decisions. Genuine human intelligence is embodied and embedded in a particular shared world and is concerned with making sense of our earthly existence in a way that enables us to enact the moves that enable all to flourish. Consequently, our human life-world will be dangerously eroded if the push for AI is allowed to continue unchecked.

We begin by discussing how a dominant paradigm of thought perpetuates the idea that human intelligence is all about abstract calculation and means-ends decision making. This framing supports the notion that more complex forms of ‘super-computation’ will enable the development of AGI. We argue that such a paradigm offers an attenuated understanding of human cognition that veils genuine human intelligence. We then attend to the way in which genuine human intellect is embodied and discursively formed within a community of others who are embedded in sustainable ways of life and who are responsive to the “slings and arrows of outrageous fortune”¹ that can derail us from time to time. At such times, our human intellect can hold us and others in being, in moments of existential uncertainly, enabling people to live well together.²

The dominant framing of human intellect

The philosopher Rene Descartes has exerted an enduring and pervasive influence on dominant western conceptions of human beings and their relation to the world.³ Descartes based his arguments on a metaphysical position that takes a reductive approach to seeking knowledge of things. This reductive approach isolates entities from their environmental context, strips them of all significance and pares them down into ‘component parts’. Descartes claimed that this process was thought to give us a sure and certain knowledge of things.

In relation to human beings, Descartes’ reductive approach opened up a chasm between the mind, the body and the world, and located the ‘human’ aspect of our Being in the mind. The mind, somehow associated with the brain, is said to be the locus of rational thought. On the Cartesian model, the ‘human’ is a fragmented and divided form of existence - human beings are a compound of a mind substance (res cogitans) and a body substance (res extensa), which somehow come together but need no other thing in order to exist. The mind, as the place of rational thought and language, is said to be an ontologically different substance from the body, and as the mind does not require the body, thought is essentially a disembodied process.⁴ To be human is to be a ‘thinking non-extended thing’ conscious of an ‘extended, non-thinking thing’, so that we are both subject and object.

The Cartesian framework sets each of us up in a private world of our own in which we establish “sure and certain knowledge” through the machinations of the rational mind. The mind, associated with the brain (and in subsequent theories that retain a Cartesian dualism, equated with the brain), is said to contain internal representations that correspond to an external reality. ‘Thinking’ is considered to be an isolated, inner process done in the mind/brain in terms of these representations and the way that they are manipulated. The ability to manipulate mental representations according to formal rules is said to be the ability to reason, and Descartes considered logic, especially mathematics, to be the quintessential form of human reasoning.⁵ As reasoning is considered to be a formal process, it is said to be universal: correct reasoning builds complex ideas out of simple ones in a way that anyone capable of applying the process can grasp. As rational beings – the archetypal animal rationale - Descartes’ epistemology claims that we are capable of expedient calculation through rigorous reflection on our ideas and mental operations. It is through this process that we can ‘know’ with absolute certainty that when we think, we exist…cogito sum.⁶

This particular approach to the framing of ‘humanity’ and ‘rationality’ forms the basis of many western practices and institutions, and it underpins the initial forays into AI and AGI.

Initial ventures into Artificial Intelligence

When human thinking is equated with the ability to calculate ‘rationally’, understood as the abstract manipulation of representations according to formal rules, one can see how those working in computer science developing ‘artificial intelligence’ envisage that we can create machines capable of surpassing humans in “all areas of reasoning”. However, when reasoning has been pared down to ‘instrumental reasoning’ or ‘calculative reasoning’, concerned with the most efficient means to a given end, meaning and significance are drained from the world.⁷ Friedrich Nietzsche was one of the first people to articulate this existential situation.⁸ He noticed that a reductive, instrumental rationality had come to dominate our approach to science, and this ‘scientific thinking’ was permeating all aspects of life. At the same time, other ways of thinking and the values and ideals they embrace, began to be seen as increasingly less ‘rational’. Over time, instrumental rationality has been increasingly adopted and normalised to the point where it has been unquestioningly and uncritically accepted. It has come to arrange and dominate all aspects of our lives, permeating our social institutions and creating bureaucratic spaces devoid of humane understanding. As a corollary, it shapes who we are and what we can do.

With ‘rationality’ unhinged from human meaning and significance, the need for ‘moral theories’ came to the fore, in order to work out the ‘right’ thing to do. Such theories, for example utilitarianism, emanate from the same impoverished mode of thinking that reduces intelligence to a calculation formula, whilst presenting it as a universal logic which it (falsely) claims is the same for everyone. In other words, instrumental reasoning has produced a way of inhabiting the world and relating to others that is hideously flawed, losing sight of the mystery and richness of what it means to be truly human.

We therefore need to pause and consider our human way of being and what has been effaced by the dominance of instrumental thinking. In so doing, we have cause to question both the approach and the outcomes of Cartesian claims, and the practices, including the development of AI and AGI, that push forward in their wake.

Our human way of being – embodied, embedded and discursively informed

When Descartes made his famous conclusion - cogito sum – he failed to ask about, or take into account, the I that thinks. For in order for any abstract, instrumental, reflective thinking to occur at all, there must be someone, somewhere, doing something that is pre-reflective; that is, there first of all has to be a self, engaged in a world of activity, who can then come to reflect upon that activity.

When we attend to the way in which human beings engage with the world, we see that human activity is first and foremost embodied activity, that is, discursively embedded in a world of human endeavours. The things we come to know are the things we come to do, and they are learnt over time through our interaction with significant others.

Our knowledge of things is through our involvement with them, and this knowledge is inscribed into our very being and it has an inexpressible aspect. For example, one can write down the rules involved with a game of football, along with explanations of player positions, the physics involved with every play of the ball, and the biomechanics of the player’s moves. But none of this grasps the way in which we learn to play football, the dispositions and qualities of character inculcated through training and participation, the reading of the nuanced bodily moves done by one’s own team and the opposition, the self-discipline required to be able to play one’s role well, and the shared aspirations and unity of purpose that makes participation meaningful.

As we learn to play the game we become attuned to the thoughts and actions of one another. We come to feel and think as others do, because “our flesh is inseparable from the flesh of the world”.⁹ In the immediacy of a unique situation we anticipate where other players are going to move next and what they are likely to do when they get there. Our understanding of what to do in each moment is shaped by cues that we are not explicitly aware of: the ‘thwack’ of the ball; the feel of the grass and soil under our feet; the movement of the wind around us. Informed by the dynamic existential feel of the situation, players come to move as one, in seamless and complex ways that embody the game and defy any linear or algorithmic programming, or calculating of ‘probability’.

Like any other human activity we are engaged in, be it making friendships to making a cup of tea, we become accomplished in that activity through being immersed in it, practicing our moves, making mistakes, being guided, corrected and encouraged by others who have been there before us, and who have our well-being in mind. At the same time we come to embody the language that frames aspects of our world, along with the dispositions and qualities of character needed to live in that world together. In the situation of coming to play football, we will be told “take it slowly, look up…. towards where you are going” and “steady….full strength now….pass to Stevie on the wing”. Such words are accompanied by the appropriate gestures and facial expressions of those who are nurturing us into this practice. In this way, through our engaged, practical involvement with things, we come to embody a range of dispositions towards the world, as we learn the meaning of a language. In this embodied, interactive way we come to be “in the world and the world is in us”.¹⁰

The complex, rich, dynamic and ineffable character of human life means that it is not possible to create a list of rules, or a programme, for all eventualities; we cannot learn a fixed set of structured dispositional responses to events in the world. Instead we must learn the art of responsively negotiating the contingencies of the world, flexibly drawing on past situations to inform our present moment in light of future considerations.¹¹ Consequently, human intelligence is a temporally integrated, hermeneutic achievement of an embodied, dispositionally informed individual embedded in a lived, meaningful environment.

As we learn to navigate the world appropriately through the guidance of others, we develop a sense of who we are, as well as what we can do and how we should do it. For example, we may make a fortnightly commitment to mow the lawns and weed the garden for an elderly neighbour, as we know that he struggles to do this job and that his family do not live close enough to help out on a regular basis. We also know how much pleasure he gets from sitting in his garden and admiring it. Our action is based on the insights we have developed over time, which orient us in light of the significance of things in this particular situation, and which illuminate what is a fitting response.¹² The motivations that inform our thinking are the values that circumscribe our shared understanding of the world. As well as being an expression of what is good in this particular situation, we also come to understand ourselves as someone who can be relied upon to uphold the values of our community. Others also come to see us in this light. Aristotle has argued that the human intellect is determined by upbringing, and not abstract calculative thought. The shaping of an individual’s reasons for acting and ways of acting are discursively constituted in community, so that her upbringing is responsible for instilling the excellences of character required to respond appropriately to the situations she encounters. That is, she perceives the situation in its entirety and responds with practical wisdom.¹³

Although we are first and foremost embodied creatures immersed in a world of pre-reflective activity, we can of course re-present and reflect on an aspect of our life and consider ways we can improve what we are doing and hence, who we become. For example, we make an inappropriate comment to a friend and see that we have upset her. We do our best to apologise, but realise that we have hurt her deeply. Reflecting on this gives us the opportunity to ask ourself such questions as: “How did I lose sight of what matters, and how can I avoid doing that again?” How can I now make amends when my verbal apology is certainly not enough? Such reflection gives us the opportunity to sharpen our insights and refine our understanding of the situation, enriching what we have to draw on as we move forward. At the same time we shape who we will become, as we consider what has happened and what we will enact next, in light of the values that we - in our community - hold to be salient and want to uphold.

What is significant to note however, is that this reflective process is a derivative mode of thinking, dependent upon having a pre-reflective understanding of things. It does not have the foundational status that the dominant Cartesian-based framework of things has given it; nor is it merely means-ends instrumental thinking, as the ends themselves are evaluated and shape our very way of being.

The move to embodied AGI

Some AI researchers have recognised the significance of our human body for thinking. Rodney Brooks is an Australian roboticist and a leading proponent of embodied cognition. Noting that experiencing the world like a human is essential to developing human-like intelligence, he argues that a machine needs to have a body — it needs to perceive, move, survive and deal with the world. Brooks also notes that abstract computation is the least important human skill, and that embodied, sensorimotor skills are essential for higher level skills like common-sense reasoning.¹⁴ Proponents of embodied cognition have also recognised the importance of engaged human interaction for enabling someone to become ‘human’. They argue that a humanoid robot makes people more comfortable in their interactions with it, and this will make it easier for the robot to learn.

However, one wonders what is fuelling the drive to create such a machine, and whether those involved with AI research fully comprehend what is entailed in our human way of being.

What narrative will an AGI machine come to inhabit, and why? Do AI researchers grasp that flesh and blood creatures such as ourselves are intimately connected to the cosmos that sustains us, and this connection must inform the narrative that we live, as our very survival depends upon it? Do they understand the way in which a unique character is informed and shaped within a discursive community, embedded within a form of life that structures a liveable narrative, framing the significance of things, and thereby the values that we enact? Do they comprehend the significance of the relationships that encourage, comfort, uphold and affirm us, so that we hold one another well?

Do they understand how our interpretation and understanding of the world, and hence, what we do, is shaped by language? For example, if we take any concept – let’s say the concept ‘bear’ - and we consider how we have come to grasp this concept through a myriad of contextual interactions which have secured its multiple roles in our conceptual system, what are we to say about this concept in a machine? Can an ‘artificial intelligence’ ever really think, as opposed to calculate? Can it know what it ‘thinks’? Do those who talk about embodied machines and push this notion of AGI really know what they are saying when they make their claims?

Conclusion

We have argued that human beings are so much more than just creatures of instrumental rationality. Although instrumental reasoning is responsible for the development of the tools, techniques and technologies that have made life easier and more comfortable for us, its dominance has come at a cost. It has arguably thinned out and devitalized the richness of our humanity, along with the world we inhabit. It has allowed the calculation of the most efficient means to a given end to replace the evaluation of those ends themselves. It has left many people bereft of understanding the significance of their life, as in a world requiring mere calculation, the self recedes; for it is when enacting practical wisdom that the self is chosen. If we continue to allow the narrow and impoverished concept of instrumental thinking and the practices that it underpins to define our world for us in an un-checked way, our humanity will suffer and our world will become less liveable.

If we uncritically allow the push for the creation of AGI to go unchallenged, we may find that the entities we meet in government departments, healthcare situations, education, lawyer’s offices etc., lack the human touch.

Dr Lynne Bowyer is an educator with a background in philosophy and mental health. Lynne has a PhD in philosophy and bioethics and has taught in primary, secondary and tertiary settings. Dr Deborah Stevens is a science communicator and educator. Her PhD in bioethics education is informed by her interdisciplinary background in science, psychology, public medicine and education.

They are founding Trustees of the Centre for Science and Citizenship (www.nzcsc.org), a charitable trust that works with students and communities throughout New Zealand promoting thoughtful engagement with the ideas and actualities that contemporary science and its accompanying technologies bring, in order to consider whether and how we can live well with them.

Endnotes

William Shakespeare, Hamlet, Act III, Scene 1.
The idea of ‘holding one another in being’ is developed in the philosophy of Martin Heidegger; see Being and Time, translated by John Macquarrie & Edward Robinson. Oxford: Basil Blackwell, 1962. It has been used beautifully by Hilde Lindemann Nelson in “What Child is This?” The Hasting’s Centre Report, 32: 6 (2002): 29-38.
René Descartes, Meditations on First Philosophy, Cambridge: Cambridge University Press, 1996, I, V.
Descartes, Meditations on First Philosophy, I, VI.
Ibid., V
Ibid., II
Max Weber, The Sociology of Religion London: Methuen, 1971, 270. Heidegger, “What calls for Thinking”, Basic Writings, from Being and Time (1927) to The Task of Thinking (1964), ed. David Farrell Krell, New York: HarperCollins, 1993,, 369-391; Theodor Adorno and Max Horkheimer, Dialectic of Enlightenment, trans. John Cumming, London: Verso, 1997, 37.
Friedrich Nietzsche, “Thus Spoke Zarathustra: A Book For All And None”, The Portable Nietzsche, ed. and trans. Walter Kaufman, New York: Penguin, 1976.
Heidegger, Being and Time, 98-99.
Heidegger, Being and Time, 376-377, 387-388.
John McDowell. Mind, Value, and Reality. Cambridge Massachusetts: Harvard University Press, 2002, 30-32.
Rodney A. Brooks & Lynn Andrea Stein. ‘Building Brains for Bodies’. Autonomous Robotics 1 (1): 7-25 (1994).
John McDowell. Mind, Value, and Reality. Cambridge Massachusetts: Harvard University Press, 2002, 30-32.
Rodney A. Brooks & Lynn Andrea Stein. ‘Building Brains for Bodies’. Autonomous Robotics 1 (1): 7-25 (1994).