AI literacy and confidence tricksters

Canada’s first AI Literacy Day was March 27.

While many forms of technology remain neutral, unbiased, objective, and impartial by their very nature, this is certainly not the case for many of the AI systems that have already been integrated into our lives. Encouraging AI literacy for people of all ages is important, as it offers space for us to seek answers to the important questions that will hopefully equip us to make informed decisions and lead to the safe and responsible use of this technology.

According to MediaSmarts, Canada’s centre for digital media literacy, AI literacy is about understanding how artificial intelligence shapes what we see and do online — and I would add, how it shapes what we see and do in the real world.

This brought to mind Ray Bradbury’s 1953 dystopian classic Fahrenheit 451. The futuristic TV parlours, with their wall-sized television screens, immersed people in shallow, chaotic noise. One could easily mistake Bradbury’s description of such rooms with that of generative AI: “It grows you any shape it wishes! It is an environment as real as the world. It becomes and is the truth.”

In other words, AI has the ability to shape the truth.

This should concern us all, especially as it relates to faith.

Faith, after all, is a lifelong journey of seeking truth through deep thought, rigorous study, intense wrestling, and attentive listening. Regardless of where these actions lead us, there is an ongoing burden of discernment that rests on our shoulders. What happens when we lose confidence in our ability to know what’s real or to believe that truth can even be found?

According to Nick Dothee, writing in The Atlantic, “If perceptual judgment starts to feel unreliable, people don’t become more analytical; they simplify, deferring to the most decisive source available, or disengage entirely.” This becomes a vicious cycle. As AI makes our judgment and discernment feel unreliable, we seek the most decisive source available, which has become AI.

Eventually, we are bound to be dizzied and thrown into disengagement.

Here’s another description Bradbury offers of the TV parlour, clearly ahead of its time: “It is immediate, it has dimension. It tells you what to think and blasts it in. It must be, right. It seems so right. It rushes you on so quickly to its own conclusions your mind hasn’t time to protest.”

A word that could be used to sum this up is ‘confidence.’ AI always speaks with total confidence. Its answers are authoritative; offering the final word on any question asked. Or as Dothee puts it: “the tone stays the same whether the answer is accurate, speculative, or completely wrong.”

Yet, has AI earned the right to such confidence on any given subject? It has not undergone the slow process of deep thought, rigorous study, intense wrestling, or attentive listening in order to arrive at any of its conclusions — but it sounds like someone who has.

While this might not seem like a big deal, the danger lies in the fact that information conveyed with confidence changes how we evaluate that information and can blur the line between fact and fiction.

This is what psychologists call the “confidence heuristic,” which Dothee describes as people’s tendency to use confidence as a shortcut for assessing credibility, even when a system is known to sometimes be wrong. This method of assessing credibility becomes elevated when evaluating life’s most difficult and challenging truths; such as questions of faith.

And even though we have already experienced the extent to which AI can “hallucinate” — offering fabricated citations, fake legal precedents, or false facts — we are not able to hold it to the same level of accountability as we do human beings.

AI does not feel the social cost of its mistakes or experience a loss of status for its errors, the way human beings might. This lack of accountability means AI simply answers the next query with the same level of confidence and we keep listening to its answers — whether true or not — until its mistakes become ours.

AI literacy can remind us that the digital world does shape the real world and online errors do lead to real life consequences. Fahrenheit 451 offers a future that looks more and more like our present.

The story cautions a future where there is no longer deep thought, rigorous study, intense wrestling, or attentive listening and the outcome is nothing more than an illusion of thought and happiness.

Whether you consider yourself a person of faith or not, process matters. Sometimes the journey is more important than the final destination.

AI literacy may be what we need to help ensure that Bradbury’s fiction doesn’t become our future.

Riley Enns is an armchair theologian and the pastor at Church of the Way.