Artificial intelligence has long been utilized to predict our preferences for movies or music, but the question remains – are people prepared to rely on AI for more serious decisions in areas like medicine or personnel selection? Recent research conducted by the University of South Australia suggests that the level of trust in AI depends on an individual’s knowledge and the significance of the decisions being made.
Surveying nearly 2000 individuals from 20 different countries, scientists discovered that when it comes to low-stakes matters such as music choices or restaurant recommendations, individuals with a strong grasp of statistics tend to trust algorithms more. However, when it involves health or employment decisions, skepticism tends to rise. Conversely, those lacking statistical knowledge and understanding of AI tend to trust algorithms across the board – from selecting a movie to receiving a medical diagnosis.
The study also found that residents of the USA, Great Britain, and Japan, as well as men and older individuals in general, exhibit more caution in placing trust in algorithms. This observation becomes increasingly significant amid the rapid proliferation of AI in various industries.
As technology advances faster than society’s comprehension of its implications, the study authors emphasize that algorithms are now playing a role not only in routine tasks but also in critical issues like healthcare and finance. This underscores the importance of understanding the limitations and potential biases of AI systems.
Experts highlight the necessity of improving AI and statistics literacy so that individuals can critically assess the reliability of algorithms. The researchers cite the case of DeepSeek, whose outcomes were influenced by the input data source, underscoring the need for transparency and reliability in algorithm development. When algorithms are built using transparent, publicly available data from reputable sources, trust in them tends to increase.