Paul Morrill
Introduction
Having spent more than thirty years in academia, I have witnessed a fair number of changes in the way universities operate, how knowledge is produced and, perhaps more importantly, how intellectual competence is recognised and rewarded. Some developments have been welcome, others rather less so. However, I cannot recall anything quite like the polarisation surrounding artificial intelligence (AI). We seem to have arrived at a point where opinion is increasingly divided between those who regard it as the beginning of something wonderful and those who appear convinced that the end is nigh.
I find myself rather uncomfortable with both positions.
This is not because I believe the truth necessarily lies somewhere between two opposing arguments. Sometimes one side is demonstrably wrong. Rather, I suspect much of the debate surrounding AI has become entangled with something considerably more personal: our identities, our accumulated abilities and the uncomfortable prospect of watching skills that once distinguished us become increasingly commonplace.
I was reminded recently of the American theologian Reinhold Niebuhr, who, in Moral Man and Immoral Society (1932, p. 221), observed that:
"Rationality belongs to the cool observers."
I have always rather liked that sentiment, although Niebuhr's original argument was more complicated than simply suggesting that detached observation produces better judgement. He recognised that human beings are not particularly good at remaining dispassionate when they have something invested in an outcome. Indeed, he went further, suggesting that those attempting to bring about social change sometimes require a degree of illusion to believe that the considerable obstacles confronting them can be overcome.
There is something rather pertinent about this when considering our responses to AI. Those who enthusiastically welcome its arrival and those who regard it with almost apocalyptic concern may have perfectly reasonable grounds for their positions, but I wonder how much of their conviction is also shaped by what they stand to gain or lose.
If pushed, I would say that I take the middle road. Not because I wish to sit on the fence, but because I think we have reached a point where understanding the technology, and perhaps our own psychological responses to it, is rather more useful than declaring allegiance to a particular camp.
Niebuhr and the problem of the cool observer
Niebuhr was writing about the relationship between individual morality, collective interests and social action. One of his central concerns was that groups and institutions frequently pursue their own interests while presenting those interests in terms of wider moral principles. Individuals may be perfectly capable of recognising the needs of others, yet the organisations to which they belong can behave rather differently.
Although his observations were made in a very different political and technological landscape, I think there is considerable relevance to the present debate.
People rarely approach important social questions without bringing their own experiences, commitments and expectations along with them. Their interpretations may therefore depend upon more than the evidence itself. Kunda's (1990) account of motivated reasoning provides a useful psychological explanation for this, suggesting that our motivations can influence how we evaluate information and the conclusions we ultimately find persuasive.
Importantly, motivated reasoning does not mean that people simply believe whatever suits them. Their preferred conclusions must still appear reasonable, which makes the process considerably more interesting. Someone with substantial knowledge of a subject may be particularly adept at constructing a convincing argument for something they already wish to believe.
I wonder how much of this is operating in the current arguments surrounding AI.
There are those who seem convinced that increasingly sophisticated systems will liberate us from routine intellectual labour, democratise knowledge and produce extraordinary advances across society. Some of these possibilities are supported by evidence. Noy and Zhang (2023), for example, demonstrated substantial productivity improvements when participants used ChatGPT to complete professional writing tasks.
However, there is a tendency among some enthusiasts to treat technological capability as though it automatically produces social improvement. That seems a rather optimistic assumption, particularly when one considers the economic and institutional interests involved.
At the opposite extreme, we have those who appear determined to regard every development as further evidence of impending disaster. Their concerns are not without foundation. Employment, intellectual property, misinformation, environmental costs and the concentration of technological power are all legitimate subjects for serious discussion.
What interests me is the conviction with which these positions are sometimes held, particularly when the consequences of accepting an opposing argument might be personally uncomfortable.
Niebuhr recognised that conviction, even when partly sustained by illusion, could be necessary for social action. Someone attempting to change an established system must believe that their efforts might succeed, perhaps despite considerable evidence to the contrary.
But the cool observer has a problem of their own. They may recognise the weaknesses in competing arguments and understand why individuals have become attached to them, but what exactly does that understanding achieve?
There is a danger of becoming so preoccupied with recognising the limitations of other people's convictions that one forgets the necessity of occasionally committing oneself to a position.
I am not entirely sure Niebuhr would have been particularly impressed by someone who spent their life observing a changing world while congratulating themselves on remaining above the argument.
Watching the dam begin to leak
My own experience of these tensions comes primarily from academia, where the development of AI has produced some rather interesting institutional responses.
Universities have spent generations constructing systems for recognising intellectual competence, awarding qualifications and distinguishing those who possess particular forms of expertise from those who do not. These arrangements have their uses, although they are certainly not without problems.
I have explored the relationship between academic authority, intellectual scarcity and AI in The Ownership of Intellect, so I will resist the temptation to revisit those arguments here.
What interests me in the present context is the almost forlorn spectacle of institutions attempting to preserve familiar arrangements while the conditions supporting them continue to change.
I cannot say that I find it particularly amusing. Having worked within universities for so long, I understand why people are anxious. There are genuine questions about academic integrity, assessment and the reliability of qualifications. Nevertheless, some of the increasingly elaborate attempts to reconstruct traditional gatekeeping structures remind me of watching someone desperately plugging holes in a dam while the pressure behind it continues to build.
Each intervention may make perfect sense to those responsible for introducing it, but I wonder whether we are spending rather too much time attempting to preserve the structure and not enough considering what might replace it.
There is also the question of what happens to people whose professional lives have been built around abilities that AI can increasingly reproduce. I considered this in I'm Sorry, Boys and Girls, but the Way Academia Currently Rewards Is Over, where I explored the relationship between academic identity, professional recognition and the possibility of experiencing a loss of occupational distinction without necessarily losing one's employment.
I have considerable sympathy for those who find these developments unsettling. Watching the gradual hollowing out of abilities once considered essential to demonstrating intellectual worth cannot be particularly pleasant, especially for someone who has spent decades developing them.
However, understanding why someone might resist technological change is not the same as demonstrating that their objections are invalid. It would be poor psychology to assume that criticism of AI must arise from professional insecurity simply because the person expressing it has something to lose.
Indeed, I suspect that accusations of psychological defensiveness have become a rather convenient way of avoiding some perfectly reasonable criticisms.
When technological progress becomes personal
My brother is an exceptionally talented special effects artist working in the film industry. He and his contemporaries possess abilities developed over many years through experience, experimentation and an extraordinary understanding of their craft.
Much of what they do involves practical knowledge that would be difficult to capture in a set of instructions. They understand how materials behave, how physical effects can be produced convincingly and how to solve problems that rarely present themselves in quite the same form twice.
It is a remarkable collection of skills, and one that has taken years to acquire.
Yet AI is beginning to occupy areas of visual production that previously required substantial specialist labour. Whether this will eventually lead to the widespread replacement of practical effects remains uncertain, and there are good reasons to believe that physical craftsmanship will continue to have a place in film production. Nevertheless, the commercial pressures are not difficult to imagine.
A studio faced with producing a convincing visual sequence may be rather more interested in the relative costs of different methods than in the years of craftsmanship required to develop them.
I can therefore understand why people working in these industries might become increasingly hostile towards AI. Imagine spending thirty years developing an ability that distinguishes you professionally, only to discover that a machine can now reproduce aspects of your work in a fraction of the time.
The ability itself has not disappeared. You have not suddenly become less accomplished, nor has your accumulated experience become worthless. But the economic value attached to those abilities may be changing, and that can have consequences extending well beyond the loss of income.
Petriglieri (2011) examined how individuals respond when important aspects of their identity appear threatened. Although her work was not concerned specifically with AI, it offers a useful way of considering why technological displacement can be experienced as something more profound than a change in employment prospects.
For someone whose working life has been organised around a particular craft, the prospect of its diminished importance may threaten their sense of purpose, achievement and professional belonging.
It seems rather unreasonable to expect those affected to greet the development with enthusiasm.
And yet there is another side to the argument, one that I find equally difficult to dismiss.
Generative AI is making forms of creative expression available to people who previously lacked the technical abilities or financial resources to realise their ideas. Someone who could never afford specialist production services may now be able to create material that would once have been entirely beyond their reach.
I regard that as a genuine benefit.
Research by Doshi and Hauser (2024) illustrates some of the difficulties in evaluating these developments. In an experiment involving short-story writing, participants given access to AI-generated ideas produced work that was evaluated more favourably, particularly among those with lower initial creativity scores. However, the resulting stories also became more similar to one another.
Their findings concern a particular form of written creativity and cannot simply be transferred to film production. Nevertheless, they demonstrate that improvements in individual creative performance can coexist with potentially undesirable consequences for collective originality.
I find myself returning to my brother's profession because it makes the argument rather less abstract than another discussion about academic prose or university assessment.
How do we celebrate the opening up of creative opportunities while acknowledging the possibility that the same technology may undermine the livelihoods of those who previously made such work possible?
I do not have a particularly satisfactory answer.
What I do know is that it is extraordinarily easy to speak enthusiastically about humanity's remarkable capacity to adapt when the costs of adaptation are being borne by somebody else.
When everyone has something to protect
There is another dimension to this debate which Niebuhr would almost certainly have recognised: the interests of the institutions involved.
Universities have understandable reasons for defending the credibility of their qualifications and the systems through which they recognise intellectual achievement. Technology companies have equally understandable commercial reasons for encouraging adoption and presenting AI as central to our collective future.
Neither position is necessarily dishonest. However, I think we would be rather naïve to assume that these interests have no influence upon how the organisations involved describe the technology.
I previously examined some of these questions in The AI Commodification of the Scalable in Higher Education, particularly the possibility that institutions may use AI to capture and reproduce aspects of staff expertise while reducing their dependence upon the individuals who developed it.
I will not rehearse that argument here, although it raises a wider question about the difference between technological possibility and institutional choice.
We are frequently told that AI is inevitable, that it will transform employment, education and creative production, and that those who fail to adapt will simply be left behind.
There may be some truth in this, but I find the argument rather too convenient when it comes from organisations that are themselves helping determine the direction of change.
A technology may be difficult to stop developing, but that does not mean every application is inevitable or that society has no authority to influence the conditions under which it is used.
Decisions concerning employment, regulation, intellectual property and the distribution of economic benefits remain matters of human judgement, however much they may be presented as unavoidable consequences of progress.
I suspect this is where some of the more enthusiastic accounts of AI become particularly unconvincing. Technological capability tells us remarkably little about whether the resulting social arrangements will be desirable.
Equally, I am reluctant to accept arguments that particular forms of expertise should remain scarce simply because their scarcity has historically provided employment or professional distinction.
The difficulty is recognising that people can have legitimate concerns while also possessing interests that influence how those concerns are expressed.
This applies as much to those promoting AI as to those resisting it, and I would be foolish to imagine that my own position is somehow exempt.
Taking the middle road
So where does this leave me?
I have described my position as taking the middle road, although I increasingly wonder whether the phrase is quite right. It suggests that I am searching for a comfortable compromise between two opposing arguments, taking something from each in the hope of arriving at a position that offends nobody.
That would be a rather disappointing ambition.
My view is simply that AI is unlikely to disappear, regardless of whether particular professions or institutions welcome its arrival. I therefore think we should learn about it, use it where appropriate, examine what it does well and remain sufficiently sceptical to recognise where its limitations matter.
I am not especially interested in defending tasks simply because they have traditionally required human effort. If a technology allows someone to accomplish something useful that was previously beyond their abilities or resources, I find that rather encouraging.
However, I am equally unwilling to accept the suggestion that every consequence of technological development must be accommodated simply because the technology exists.
This is where Niebuhr's observations become particularly interesting.
The committed opponent of AI may occasionally overstate the dangers, but their resistance could nevertheless produce worthwhile outcomes. Without people challenging technological developments, questions concerning employment protection, creative ownership and the concentration of corporate power might receive considerably less attention.
The enthusiast may be excessively optimistic, yet that optimism could encourage experimentation and produce benefits that a more cautious observer might never have pursued.
I do not think acknowledging these possibilities requires us to regard both positions as equally convincing. It merely suggests that social progress may depend upon people whose convictions are stronger than the available evidence entirely warrants.
And what of the cool observer?
There is a danger that continually recognising the limitations of competing positions becomes a form of intellectual paralysis. One can become so adept at explaining why everyone else is mistaken that one never quite gets around to doing anything oneself.
I would not want my middle road to become an excuse for that.
Nor would I wish to portray myself as some uniquely dispassionate observer. I have spent much of my adult life in academia, and I possess many of the abilities whose value is now being questioned. My enthusiasm for AI's potential to broaden intellectual participation is also a position influenced by my own experiences and preferences.
There is no particular reason to suppose that I am immune to the same psychological processes I recognise in others.
Perhaps the most we can reasonably ask of ourselves is to remain aware of those influences, examine evidence carefully and retain a willingness to reconsider our position when circumstances change.
Conclusion
I began by suggesting that our responses to AI may reveal something interesting about human psychology, particularly when technological change begins to challenge abilities and arrangements around which people have organised their lives.
For some, AI represents an opportunity to participate in intellectual or creative activities from which they were previously excluded. For others, it threatens skills acquired over decades, together with the livelihoods and professional identities those skills have helped sustain.
I find little pleasure in watching people struggle to preserve something into which they have invested so much of their lives. Equally, I find it difficult to defend established boundaries simply because those within them have become accustomed to the advantages they provide.
We have adapted to technological change before, although I think we should be careful about treating that history as evidence that the process will be painless or universally beneficial. Adaptation can be extraordinarily costly, particularly for individuals who have little influence over the changes being imposed upon them.
My position, for what it is worth, is that we should engage with AI as a technology that is already changing the world around us, while remaining prepared to challenge the ways in which it is developed and deployed.
I also think we could do with rather less certainty on all sides.
After more than thirty years in academia, I have become increasingly suspicious of people who appear convinced that they possess the definitive interpretation of a complicated problem. I am particularly suspicious when that interpretation happens to align rather conveniently with their professional or economic interests.
Of course, I have no guarantee that my own interpretation is any more reliable.
Niebuhr's cool observer is an appealing figure, but perhaps the difficult part is recognising when observation must give way to action, without allowing conviction to become a substitute for judgement.
I would like to think that taking the middle road means remaining open to technological possibilities without becoming an evangelist for them, and recognising genuine human losses without concluding that the world should therefore remain as it was.
Whether that is a sustainable position remains to be seen.
I may, of course, be completely wrong about all of this.
After thirty years in academia, it would hardly be the first time.
References
Doshi, A. R., & Hauser, O. P. (2024). Generative AI enhances individual creativity but reduces the collective diversity of novel content. Science Advances, 10(28), eadn5290. https://doi.org/10.1126/sciadv.adn5290
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108(3), 480–498. https://doi.org/10.1037/0033-2909.108.3.480
Niebuhr, R. (1932). Moral man and immoral society: A study in ethics and politics. Charles Scribner's Sons.
Noy, S., & Zhang, W. (2023). Experimental evidence on the productivity effects of generative artificial intelligence. Science, 381(6654), 187–192. https://doi.org/10.1126/science.adh2586
Petriglieri, J. L. (2011). Under threat: Responses to and the consequences of threats to individuals' identities. Academy of Management Review, 36(4), 641–662. https://doi.org/10.5465/amr.2009.0087