In uncertain times, religion can provide a sense of clarity about a person’s life or ultimate destination, and where they fit into the world. However, societies include diverse religions and many people who have none, which can bring uncertainty about what’s true. This can be reduced by joining a group or holding a worldview that leans on a certain and benevolent authority. While it may not last, this can be alluring.
At the same time for some people artificial intelligence, as it is often portrayed in media, can encourage reliance on its authority, perhaps with some fear when it is portrayed as transcending human limitations. And in some ways like religion, this authority can be made approachable through certain acts or rituals. Indeed, this accessibility is heightened when media portrayals of AI focus on generative text chat models since they present human-like qualities and are often portrayed as anthropomorphised robots – beautiful, androgynous, and intelligent.
This can lead people to talk about AI algorithms as agents with their own identity, leaning into their responses and feeling thankful for their recommendations. This does not account for the human labour at the heart of the datasets that shape and train AI models and the computer code that produces these outputs. Nor does it account for how AI algorithms’ outputs are probabilistic or contingent on certain premises, which may not be clear. Of course, marketers can lean into this perception of intelligence and even infallibility when promoting AI products and services and besides, cultivating certainty and reliability can yield sales.
Researchers can raise critical awareness of the realities of AI but they need to know how the coding that underlies these algorithms works. A (perhaps) small but increasing proportion of non-computer science scholars have critical understanding at this intersection and I hope to contribute through my research. This perspective will be essential to helping people remain realistic about how AI and machine learning can benefit their lives while retaining the vulnerability of knowing our machines don’t have all the answers to life’s questions.
Tim Kinnear 