AI and the humanities: crisis or opportunity?
This blog was kindly authored by Scarlett Keating, a Year 12 student who recently completed work experience at HEPI.
In 1965, Cambridge History professor JH Plumb’s edit of ‘The crisis in humanities’ was published, citing the struggle of the humanities to stay relevant in a world of rapid technological and scientific growth. While Plumb was incorrect to imply a collapse of humanities study, he rightly highlighted the need for academic humanists to adapt to a tech-driven world.
Today, the humanities face a similar crisis: the expansion of Large Language Model (LLM) use poses a potential threat to the value of human critical thinking and understanding, while a blog by LSE by Marvin Starominski-Uehara highlights a further risk: devaluing university degrees. If students are submitting AI-generated work, and professors are unable to recognise it as AI (one recent study states that, across eight universities, professors correctly identified AI papers only 53% of the time), then students will graduate without all the skills supposedly gained from their degree. Upon entering employment, this gap will be revealed, which risks discrediting qualifications for all graduates, not just those who (mis)used AI.
AI can serve useful purposes in the study of the humanities which do not impede human skills: generating basic summaries, translating collections of historical text, and managing bibliographies are all practical uses of AI as an assistive tool. However, understanding, argument building, and interpretation of ethics (all areas where AI falls short of human capabilities) should not be short circuited.
While the HEPI / Kortext Student Generative AI Survey found only 12% of respondents placed AI-generated text directly in their assessed work (and a further 25% used edited AI generated text), even this could represent a sufficiently large proportion of students to cause reputational harm for all. However, this issue may also provide the solution to the humanities’ perceived growing irrelevance: graduates who have not supplemented their learning with AI may now present even more value to employers.
So, is there a crisis in humanities, or does the crisis make a humanities degree even more important?
The growth of AI presents a challenge to higher education institutions. They need to guide students to enhance their uniquely human understanding of ideas and ethics. If some people are graduating without the skills certified by their qualification due to misuse of AI, this provides an opportunity for students who have developed inherently human skills through their studies. While AI can diminish the abilities of those who use it excessively and improperly, genuine and interested humanities students face limited competition from AI. This is particularly the case when it comes to the interpretation of ideas, empathy and ethics and the ability to understand ideas alongside cultural nuances and temporal context – largely due to the lack of development in these areas of AI function. This creates a significant disparity between the quality of graduates, making those who truly have the expertise AI lacks stronger candidates for employers and further study opportunities.
Study of the humanities allows students to excel in the areas where AI is lacking for a number of reasons. The humanities nurture an understanding of ethics and the ability to evaluate moral consequences; the ability to think critically by judging the credibility of sources and arguments; and developing an understanding of human behaviours and contexts which can be applied to judgements, while subjects which champion the use of data and calculations may struggle to hold an advantage over AI.
In part, the responsibility for finding the right skills lies with employers, who should ensure their application processes are set up to root out applicants who have graduated with gaps in their capabilities: writing tasks, source interpretation or live problem solving, for example, test authentic human skills. However, it is also vital that universities work to ensure their graduates have high levels of ability in areas where humans excel beyond the scope of AI replication.
There is a head of steam behind the idea that we need assessment methods that better assess human judgement and thinking, as recent pieces from both the University of Manchester and HEPI shows. These include assessments where it is impossible to rely entirely on AI – such as oral exams or asking people to apply what they know to applications to real-world problems. Not only would this prevent the abuse of AI in assessments and coursework, but it would ensure that all students are stretched and taught in ways which speak to their human aspects. This will produce graduates better adapted to a world of AI, and better equipped to surpass competition for jobs and research.
On top of this, in a world dominated by artificial intelligence and technological advancements, it is more important than ever to safeguard the uniquely human qualities which define the humanities. Institutions which have long served as havens for the research and study of the humanities must strive to protect these subjects, and their graduates, from irrelevance.
Comments
Paul Vincent Smith says:
I hope the author doesn’t find it patronising if I say what an interesting and readable piece this is.
There are of course a few stand-out issues, as there inevitably are when we speak of AI and HE, particularly HE assessment. For a long time I have advocated dual summative assessment, that is assessment optionality, for all HE course units: one that is “live” (prevents AI support at the time of doing the assessment) and one that is “take-home” (we know that AI might be used in the completion of it). These choices would be recorded on transcripts. How to choose graduates therefore becomes a moral as well as a practical decision for employers; it also presents students with their own choices. This proposed offsets the dilemmas faced by universities, which they/we seem to be facing unilaterally. I think it more than reasonable to share these dilemmas with those who are implicated in the wider AI and employment ecosystem.
I do not see employers rushing to make AI use in assessment and degree outcomes a moral question if it is not placed at their door. Similarly, although we hear a lot about the requirement of employers for “AI skills”, I have still not seen any compelling idea as to what these generic skills are, and there is research that suggests employers struggle to name them. The research from my own university cited here is positive inasmuch as it promotes “competencies” rather than “skills” – a subtlety that appears to have escaped the PR team. Let us talk about the meaning of AI and support discussion about its application and implications; employers can train graduates to use such platforms in the workplace.
Reply
Jonathan Alltimes says:
What are the humanities?
Lumping a whole load of disparate specialisms together was an error and the cause of the ongoing dysfunctions. The idea of the humanities is a misnomer if ever I saw one. AI has shown us that we do not know what we are doing. The elites studied Latin so they could communicate in a common language for various purposes, it was language of cohesion in the state, the military, administration, and trade. The texts became the substrate for preserving the learning of the language and later as the Byzantium Empire declined and passed itself to later European empires, the texts became relics for preserving what was once powerful and so it was thought could retain ideas about being powerful. The idea of a crisis is the disequilibrium of a system and its regulation. It is a model for scientific explanation borrowed from ancient Greek medical practice and the body, which relies on the principle of the bipolarity of opposites, for example, hot and cold in fevers and their epidemics. Academics and students need a hinterland of experiences with which to use AI. Can you estimate?
Reply
Add comment