Markus Luczak-Roesch

Ronald Fischer, Markus Luczak-Roesch, Johannes A Karl

There has been concern about ideological basis and possible discrimination in text generated by Large Language Models (LLMs). We test possible value biases in ChatGPT using a psychological value theory. We designed a simple experiment in which we used a number of different probes derived from the Schwartz basic value theory (items from the revised Portrait Value Questionnaire, the value type definitions, value names). We prompted ChatGPT via the OpenAI API repeatedly to generate text and then analyzed the generated corpus for value content with a theory-driven value dictionary using a bag of words approach. Overall, we found little evidence of explicit value bias. The results showed sufficient construct and discriminant validity for the generated text in line with the theoretical predictions of the psychological model, which suggests that the value content was carried through into the outputs with high fidelity. We saw some merging of socially oriented values, which may suggest that these values are less clearly differentiated at a linguistic level or alternatively, this mixing may reflect underlying universal human motivations. We outline some possible applications of our findings for both applications of ChatGPT for corporate usage and policy making as well as future research avenues. We also highlight possible implications of this relatively high-fidelity replication of motivational content using a linguistic model for the theorizing about human values.

Cameron Lai, Markus Luczak-Roesch

The influence of Web search personalisation on professional knowledge work is an understudied area. Here we investigate how public sector officials self-assess their dependency on the Google Web search engine, whether they are aware of the potential impact of algorithmic biases on their ability to retrieve all relevant information, and how much relevant information may actually be missed due to Web search personalisation. We find that the majority of participants in our experimental study are neither aware that there is a potential problem nor do they have a strategy to mitigate the risk of missing relevant information when performing online searches. Most significantly, we provide empirical evidence that up to 20% of relevant information may be missed due to Web search personalisation. This work has significant implications for Web research by public sector professionals, who should be provided with training about the potential algorithmic biases that may affect their judgments and decision making, as well as clear guidelines how to minimise the risk of missing relevant information.

Cathal Doyle, Yevgeniya Li, Markus Luczak-Roesch et al.

In this paper we seek to contribute to the debate about the nature of citizen involvement in real scientific projects by the means of online tools that facilitate crowdsourcing and collaboration. We focus on an understudied area, the impact of online citizen science participation on the science education of school age children. We present a binary tree of online citizen science process flows and the results of an anonymous survey among primary school teachers in New Zealand that are known advocates of science education. Our findings reveal why teachers are interested in using online citizen science in classroom activities and what they are looking for when making their choice for a particular project to use. From these characteristics we derive recommendations for the optimal embedding of online citizen science in education related to the process, the context, and the dissemination of results.

Markus Luczak-Roesch

Location-based augmented reality games have entered the mainstream with the nearly overnight success of Niantic's Pokémon Go. Unlike traditional video games, the fact that players of such games carry out actions in the external, physical world to accomplish in-game objectives means that the large-scale adoption of such games motivate people, en masse, to do things and go places they would not have otherwise done in unprecedented ways. The social implications of such mass-mobilisation of individual players are, in general, difficult to anticipate or characterise, even for the short-term. In this work, we focus on disaster relief, and the short- and long-term implications that a proliferation of AR games like Pokémon Go, may have in disaster-prone regions of the world. We take a distributed cognition approach and focus on one natural disaster-prone region of New Zealand, the city of Wellington.