A new study suggests that increasing human involvement in the deployment of AI for public services can positively impact the acceptance of this technology.
According to TechXplore, the study highlights that citizens' apprehensions go beyond just AI fairness, encompassing potential biases that may arise from human inputs.
When administrative discretion is perceived as overly extensive, citizens tend to lean towards embracing AI usage, according to the study.
The research was conducted by Laszlo Horvath from Birkbeck, University of London, in collaboration with Oliver James, Susan Banducci, and Ana Beduschi from the University of Exeter.
AI Involvement
In a survey involving 2,143 participants in the UK, individuals were tasked with indicating their preferences regarding the extent of AI involvement in the processes associated with immigration visa and parking permit handling.
The findings indicated that increased human involvement generally correlated with a greater acceptance of AI. However, in cases where there was significant human discretion involved in parking permit processing, respondents were inclined towards favoring more limited human intervention.
System-level elements, such as high accuracy, the availability of an appeals process, heightened transparency, reduced costs, non-sharing of data, and the absence of private company involvement, all contributed to an elevated level of acceptance and a perceived sense of procedural fairness.
Horvath said that their findings suggest a reluctance towards the accumulation and sharing of citizens' data. However, in the context of other characteristics at the system level, they observed that citizens desire functional technology and, in such cases, are open to a reduced level of human oversight.
Banducci added that the results contribute to the understanding of technology acceptance in digital government and AI. Citizens who may initially resist new technologies in several contexts appear to prefer increased human administrative involvement.
Read Also: OpenAI's New 'Preparedness' Team Focuses on Countering 'Catastrophic' AI Risks, 'Human Extinction'
Relevance to Government Services
James underscored that their findings have broad relevance to government services employing AI, as many routine interactions with the government involve permit applications similar to those they scrutinized.
"Many routine interactions with the government involve permit applications similar to the kinds we examined such that the findings are of broad relevance to government services using AI," James said in a press statement.
"Respondents appeared to be more strongly influenced by the costs and accuracy of the technology than by concerns about 'humans in the loop,' transparency, or even data sharing. This suggests citizens may perceive legitimacy more profoundly in terms of the system's efficiency and its ability to deliver accurate and cost-effective results," he added.
The study holds relevance amid the ongoing concerns surrounding AI due to its potential to address critical issues. It provides valuable insights into how citizens perceive and accept AI in public services, shedding light on the importance of human involvement and system-level factors.
The findings of the team were published in Government Information Quarterly.
Related Article: OpenAI to Launch Major Updates to Make AI Software Development Cheaper and Faster, Luring Developers
