ChatGPT and the Generation of Digitally Born “Knowledge”: How Does a Generative AI Language Model Interpret Cultural Heritage Values?

Abstract

The public release of ChatGPT, a generative artificial intelligence language model, caused wide-spread public interest in its abilities but also concern about the implications of the application on academia, depending on whether it was deemed benevolent (e.g., supporting analysis and simplification of tasks) or malevolent (e.g., assignment writing and academic misconduct). While ChatGPT has been shown to provide answers of sufficient quality to pass some university exams, its capacity to write essays that require an exploration of value concepts is unknown. This paper presents the results of a study where ChatGPT-4 (released May 2023) was tasked with writing a 1500-word essay to discuss the nature of values used in the assessment of cultural heritage significance. Based on an analysis of 36 iterations, ChatGPT wrote essays of limited length with about 50% of the stipulated word count being primarily descriptive and without any depth or complexity. The concepts, which are often flawed and suffer from inverted logic, are presented in an arbitrary sequence with limited coherence and without any defined line of argument. Given that it is a generative language model, ChatGPT often splits concepts and uses one or more words to develop tangential arguments. While ChatGPT provides references as tasked, many are fictitious, albeit with plausible authors and titles. At present, ChatGPT has the ability to critique its own work but seems unable to incorporate that critique in a meaningful way to improve a previous draft. Setting aside conceptual flaws such as inverted logic, several of the essays could possibly pass as a junior high school assignment but fall short of what would be expected in senior school, let alone at a college or university level.

The Role of Artificial Intelligence in the Study of the Psychology of Religion

“…The utilization of artificial intelligence (AI) for the analysis of religious texts presents several challenges and limitations, primarily in relation to the cultural context and linguistic nuances inherent in such texts (Spennemann 2023). The application of AI techniques in this domain requires careful consideration of these factors (Grassini 2023).…”

Will Artificial Intelligence Affect How Cultural Heritage Will Be Managed in the Future? Responses Generated by Four genAI Models

“…At present, genAI language models rely on statistical connections and patterns within their training data. At the time of writing, they lack reasoning ability [23] and do not possess independent creative thinking abilities, which makes them susceptible to inverted logic phenomena [24]. Despite being capable of producing complex text predictions, any demonstrated creativity in these models is solely dependent on the perception of the human interacting with the language model.…”

“…The dataset underpinning ChatGPT 3.5, which also relies on 175 billion parameters, had a data cut-off in September 2021, which means they cannot accommodate or comment on events, discoveries, or perspectives that have emerged after this date (Table 1). ChatGPT can, to some degree, incorporate in its responses the context of or prior answers in the same chat and can critique its own response, but is unable to then revise that initial response in a meaningful way [24]. The nature of the training data is not publicly available, with OpenAI noting that ChatGPT 3.5 used "three primary sources of information: (1) information that is publicly available on the internet, (2) information that we license from third parties, and (3) information that our users or our human trainers provide" [30].…”

“…Additionally, in a previous paper, the present author looked at how genAI language models can present a cohesive discussion of values in cultural heritage management [24]. While genAI language models have been shown to be competent at extracting, summarising, and integrating information, they are less suitable (at least in their present configuration) at presenting a critical discussion, as they may suffer from inverted logic [24].…”

“…Previous work had shown that genAI language models are, on occasion, prone to confabulation (or 'hallucinations'), i.e., responses where erroneous 'facts' are being offered [97,98], where answers are provided with inverted logic [24], or where references are invented [5,24]. This confabulation is caused when the predictive contextual modelling goes off on a tangent and then follows that line of 'thought' without reverifying logical coherence [24]. Examples for these were offered both by ChatGPT and DeepAI.…”