Large Language Models (LLMs) are expected to significantly impact various socio-technical systems, offering transformative possibilities for improved interaction between humans and technology. However, their integration poses complex challenges due to the intricate interplay between societal structures, human behaviour, and technological innovation. This research explores these multifaceted challenges, emphasising the need for a human-centered approach in integrating LLMs to ensure that technological advancements are aligned with ethical standards and societal needs. Utilizing a structured methodology comprising a workshop, literature analysis, and expert collaborations, the study uses a multi-dimensional human-centered AI framework to guide the responsible integration of LLMs. Key insights include the importance of inclusive data, considering unintended consequences, maintaining privacy, and respecting intellectual property rights. The paper identifies and advocates for principles like human-in-the-loop, continuous longitudinal studies, proactive awareness campaigns, and regular audits to develop LLMs that are ethically sound, adaptable, and effectively integrated into various socio-technical systems, thus addressing user needs and broader societal impacts. The paper also underlines the importance of collaboration among academia, industry, and policymakers to develop LLMs that are ethically aligned, socially beneficial, and adaptable to future societal needs. The findings offer valuable insights into the strategic integration of LLMs, advocating for a broader research perspective beyond industrial motivations to fully understand and leverage LLMs in socio-technical landscapes.

@en

ChatGPT is a highly advanced AI language model that has gained widespread popularity. It is trained to understand and generate human language and is used in various applications, including automated customer service, chatbots, and content generation. While it has the potential to offer many benefits, there are also concerns about its potential for misuse, particularly in relation to providing inappropriate or harmful safety-related information. To explore ChatGPT's (specifically version 3.5) capabilities in providing safety-related advice, a multidisciplinary consortium of experts was formed to analyse nine cases across different safety domains: using mobile phones while driving, supervising children around water, crowd management guidelines, precautions to prevent falls in older people, air pollution when exercising, intervening when a colleague is distressed, managing job demands to prevent burnout, protecting personal data in fitness apps, and fatigue when operating heavy machinery. The experts concluded that there is potential for significant risks when using ChatGPT as a source of information and advice for safety-related issues. ChatGPT provided incorrect or potentially harmful statements and emphasised individual responsibility, potentially leading to ecological fallacy. The study highlights the need for caution when using ChatGPT for safety-related information and expert verification, as well as the need for ethical considerations and safeguards to ensure users understand the limitations and receive appropriate advice, especially in low- and middle-income countries. The results of this investigation serve as a reminder that while AI technology continues to advance, caution must be exercised to ensure that its applications do not pose a threat to public safety.@en

Participatory data stewardship (PDS) empowers individuals to shape and govern their data via responsible collection and use. As artificial intelligence (AI) requires massive amounts of data, research must assess what factors predict consumers’ willingness to provide their data to AI. This mixed-methods study applied the extended Technology Acceptance Model (TAM) with additional predictors of trust and subjective norms. Participants’ data donation profile was also measured to assess the influence of individuals’ social duty, understanding of the purpose and guilt. Participants (N = 322) completed an experimental survey. Individuals were willing to provide data to AI via PDS when they believed it was their social duty, understood the purpose and trusted AI. However, the TAM may not be a complete model for assessing user willingness. This study establishes that individuals value the importance of trusting and comprehending the broader societal impact of AI when providing their data to AI. Practitioner summary: To build responsible and representative AI, individuals are needed to participate in data stewardship. The factors driving willingness to participate in such methods were studied via an online survey. Trust, social duty and understanding the purpose significantly predicted willingness to provide data to AI via participatory data stewardship.

@en

Artificial Intelligence (AI) agents are predicted to infiltrate most industries within the next decade, creating a personal, industrial, and social shift towards the new technology. As a result, there has been a surge of interest and research towards user acceptance of AI technology in recent years. However, the existing research appears dispersed and lacks systematic synthesis, limiting our understanding of user acceptance of AI technologies. To address this gap in the literature, we conducted a systematic review following the Preferred Reporting Items for Systematic Reviews and meta-Analysis guidelines using five databases: EBSCO host, Embase, Inspec (Engineering Village host), Scopus, and Web of Science. Papers were required to focus on both user acceptance and AI technology. Acceptance was defined as the behavioural intention or willingness to use, buy, or try a good or service. A total of 7912 articles were identified in the database search. Sixty articles were included in the review. Most studies (n = 31) did not define AI in their papers, and 38 studies did not define AI for their participants. The extended Technology Acceptance Model (TAM) was the most frequently used theory to assess user acceptance of AI technologies. Perceived usefulness, performance expectancy, attitudes, trust, and effort expectancy significantly and positively predicted behavioural intention, willingness, and use behaviour of AI across multiple industries. However, in some cultural scenarios, it appears that the need for human contact cannot be replicated or replaced by AI, no matter the perceived usefulness or perceived ease of use. Given that most of the methodological approaches present in the literature have relied on self-reported data, further research using naturalistic methods is needed to validate the theoretical model/s that best predict the adoption of AI technologies.@en

Artificial intelligence (AI) chatbots are set to be the defining technology of the next decade due to their ability to increase human capability at a low cost. However, more research is required to assess individuals' behavioural intentions to use this technology when it becomes publicly available. This study applied an extended Technology Acceptance Model (TAM), with additional predictors of trust and privacy concerns, to assess individuals' behavioural intentions to use AI chatbots across three industries: mental health care, online shopping, and online banking. These services were selected due to the current popularity of regular chatbots in these fields. Participants (N=360, 202 females) aged between 17 and 85 years (M=38.17, SD=17.66) completed a 71-item online, cross-sectional survey. As hypothesised, perceived usefulness and trust were significant positive predictors of behavioural intentions across all three behaviours. However, the influence of the perceived ease of use and privacy concerns on behavioural intentions differed across the three behaviours. These findings highlight that the combination of predictors within the extended TAM have different influences on behavioural intentions to use AI chatbots for mental health care, online shopping, and online banking. This research contributes to the literature by demonstrating that the influence of the variables in one field cannot be generalised across all uses of AI chatbots.

@en