As AI begins to reshape entry-level employment in tech-heavy fields, new research suggests a counterintuitive safe haven in an unlikely place: cultural heritage.
A recent study from Stanford University researchers, Canaries in the Coal Mine?, finds that early-career workers (ages 22–25) in AI-exposed occupations, such as software engineering and customer support, have experienced a 13% relative decline in employment. In contrast, employment for older workers in these same roles has remained stable or grown, suggesting that AI is eating into codified, ‘textbook’ tasks while preserving or even rewarding experience and tacit knowledge.
This shift raises an urgent question: what kinds of careers still offer entry points where judgment, context, and cumulative knowledge remain the core currency?
Here’s where the human, so-called “soft”, skills in cultural heritage may have an edge.
A Bastion of Tacit Knowledge
Unlike AI-exposed professions built around repeatable problem-solving or scripted interactions, heritage roles — from conservation and collections management to oral history and public engagement — thrive on the nuanced, the situated, and the local. These are contexts where AI can provide useful tools but rarely offers true substitution.
Take for instance the complexities of managing a listed building, interpreting a contested historical narrative, or facilitating inclusive community participation in heritage preservation. These are not tasks easily handed over to generative AI. They demand historical judgment, emotional intelligence, and place-based understanding — what economists like Paul David [1] and Eric von Hippel [2] have called “sticky knowledge,” and what educational theorists, like Schön [3] and Lave and Wenger [4] frame as tacit, experiential, and situated learning. It’s exactly this type of expertise the Stanford paper shows is resilient to AI displacement.
While the AI-driven labour market is hollowing out junior tech roles, cultural heritage remains undervalued and underfunded. In the UK, routes into heritage work are often bottlenecked by unpaid internships, patchy or siloed investment in skills development, and limited diversity in recruitment. Despite this, heritage workers contribute around £67,000 to the economy per head, comparable to car manufacturing or film production. Yet the cultural heritage sector continues to face long-standing workforce planning challenges, including the loss of experienced staff without structured succession or upskilling pathways (Prospect, 2023).
This labour market mismatch may soon flip. As industries reliant on AI automation trim junior roles, the heritage sector offers something rare: a domain still hungry for human interpretation, intergenerational learning, and localised expertise. With the right interventions, for example, funded apprenticeships or practice-based degrees, cultural work could absorb some of the talent displaced by AI.
Proof of Polanyi’s Paradox
The Stanford paper draws a critical distinction between AI that augments work and AI that automates it. Employment declines appear only in the latter. In heritage, AI can help with cataloguing archives or digitising artefacts, but it can’t lead a school visit in real life (IRL) with the same humour or emotion as a human guide. And philosophical critiques, such as those rooted in Polanyi’s paradox [5], remind us that AI remains ill-equipped to handle the tacit, intuitive knowledge needed to co-create projects with older or migrant communities or make nuanced judgement calls about climate impacts on fragile sites.
This positions cultural work not only as resistant to automation, but potentially enhanced by responsible AI use if adopted ethically. It also means cultural careers could become a model for balanced augmentation, where tools support professionals rather than replace them.
A Wake-Up Call for Heritage Skills Policy
For this to happen, cultural institutions and funders must stop treating heritage careers as vocational afterthoughts and start recognising them as vital civic infrastructure for the AI age. That means accrediting experiential learning, and reducing degree inflation for roles that rely on site-based skills rather than abstract knowledge work.
In light of Brynjolfsson and colleagues’ findings, we should ask why heritage roles haven’t already been rebranded as resilient, future-proof options, especially for fresh graduates and less experienced workers seeking meaningful, place-rooted work in a volatile labour market.
The opportunity? To position heritage careers as a 21st-century “tacit economy,” a space where AI is unlikely to replace human insight, and where experience can be leveraged rather than bypassed.
Because what if, instead of being the canaries in the coal mine, cultural workers became the guardians of reality and authenticity in a virtual world — custodians of meaning, memory, and judgment, holding the line where algorithms end and human interpretation begins.
If you’re working at the intersection of heritage, learning, or AI, I’d love to know: What strategies are you seeing, or wishing for, to make cultural work more visible, viable, and valued?
References Worth Reading:
[1] David, P.A. (1993). Knowledge, property, and the system dynamics of technological change. Proceedings of the World Bank Annual Conference on Development Economics, 1992, pp. 215–248.
[2] von Hippel, E. (1994). “Sticky information” and the locus of problem solving: Implications for innovation. Management Science, 40(4), pp. 429–439.
[3] Schön, D.A. (1983). The Reflective Practitioner: How Professionals Think in Action. New York: Basic Books.
[4] Lave, J. and Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press.
[5] Polanyi, M. (1966). The Tacit Dimension. London: Routledge & Kegan Paul.
This article was written for and first published on Access Heritage.
Leave a Reply