Serious conversations

Earlier this month, Pope Leo XIV issued a 42,000-word encyclical on AI and the challenges it poses to humanity. I believe it is a wonderful thing that faith leaders are sharing their perspective. The Pope, who has an undergraduate degree in mathematics, has spoken about AI before, and the encyclical is an important and timely document. It is rich in detail and offers us an opportunity to reflect about our present and the future.

As we engage with generative aspects of AI, there are important questions we too must ask ourselves to have a morally and intellectually consistent framework of our lives. For example, how do those - who care deeply about the environment and often express their frustration at how policies of the rich countries are causing global harm disproportionately affecting the poor countries - square the environmental impact of data centres at the core of AI growth?

Are those who feel strongly about protection of intellectual property, especially of artists and writers and from low- and middle-income countries, okay with widespread copyright infringement and acquisition of creative works without consent by large AI companies?

There are others who have deep concerns about the overreach of global powers and the diminishing rights of the weak. But as tech companies start to dominate global capital, influence policy and bypass regulations, how do they feel about full-scale indulgence in these products?

I also wonder about teachers who choose the noble profession not because of high salaries but the deep and lasting satisfaction of training the next generation, of helping people become more sophisticated in their original writing, analysis and thought - encouraging their students to look for a quick answer all in the name of efficiency.

For many of us in research, the AI tools to summarise a paper or get to the core argument of a study are deeply unsatisfactory. It may be fast, but it is often inaccurate and routinely fails to pick nuance. I wonder if our goal in research is speed, and not accuracy.

I recognise there are strong arguments to be made on the other side. There is evidence of improved computation, a sharper analysis of a diagnostic image to find some early signs of a serious ailment, the ability to search for new therapeutic molecules, and to design experiments that were previously not thought of. But that is hardly what is being promoted from the top. With the exception of a small group, new mathematical insights are not what is being discussed. When people talk about things changing rapidly, they often talk about doing less creative work, reading fewer pages, having fewer workers doing the hard work, skimming the final draft and mostly not writing anything. In my circle of professors, colleagues brag about having an AI agent write letters of recommendation for their students. The professors never tell the students that the letters were not written by them. In some cases, they do not even bother to proofread them for basic errors. The students may not get the job or the admission, but no one holds the letter writer responsible for a bad letter. What most people are excited about is some notion of efficiency, in pursuit of material gains. That pursuit will surely bring some gains and will come at a cost to the environment, creative and cognitive skills and our relationship with each other - and those costs need to be discussed with seriousness and not hyperbole.

The question of humanity, truth and honesty are central to any faith. And I hope that faith leaders from all traditions will also offer both their critique and their reflections. Faith leaders are an important group that should weigh in with rigorous arguments. But so should others who care about what it means to be human. If we are to believe that AI is touching all aspects of our lives, then serious people from all sectors need to reflect on what this new technology means for them, and how to navigate, engage or resist rapid changes. No technology or innovation is inevitable for as long as we believe in our agency.