On May 15, 2026, Pope Leo XIV published his first encyclical, Magnifica Humanitas. This is a letter to the Catholic Church addressing the rising use of AI and how the Church should respond. He criticizes the unethical and harmful use of AI, especially in war. Observing the increasing capabilities of autonomous weapons, he condemns the march towards machines that can operate without human oversight and the destruction they may cause. On a broader scale, he fears that AI is slowly replacing human agency in society, and that if we continue on our current path, we’ll be “constructing Babel”. The Pope is ultimately concerned with “preserving the human person in the age of artificial intelligence,” which he fears is at risk as people become evermore reliant on technology.
These comments, coming from the head of a religious institution with over 1.4 billion members, represent a broader dissatisfaction with AI. Like Pope Leo, many people fear the negative effects on society and what the future may look like. The public often cites environmental, economic, and social concerns regarding how AI will change individuals and communities going forward. They see headlines like the global AI footprint being estimated at 4 to 6 billion gallons of water a year, Meta letting go of 8,000 employees as it seeks to increase its use and spending on AI by billions, or fears the AI bubble could be four times as bad as its dot-com counterpart. When former Google CEO Eric Schmidt spoke at the commencement at the University of Arizona promoting AI, he received boos and jeers. Regardless of the future and how AI will be involved in it, the optics surrounding it right now are, frankly, terrible. Often, the only places you’ll find people excited about AI are specific communities like on Reddit, Twitter/X, or YouTube, along with companies seeking to improve their efficiency.
However, I, myself, have always been fascinated by AI and its potential. I heard stories and predictions about the bountiful, effectively utopian future brought about by this technology at the advent of artificial general intelligence (AGI), a machine that can perform at the same level or higher as a human in all intellectual areas, if not also physical. Diseases would be cured, the environment would be restored, and poverty and starvation would be eliminated through advanced prediction models acting as oracles, nanobots killing individual harmful bacteria and viruses as they were forming. Evidently, this is not the view or hope shared by most. But I think I have a way to change that.
There are, in fact, multiple AI products that have been well-received by the public. These tools aim to solve specific problems and increase the efficiency of tasks. One of the best examples of this is NotebookLM, which is Google’s research and note-taking tool that summarizes the sources you give it. People use it to decrease the time learning about a topic, organize an essay outline with specific sources for each section, and find trends or generate ideas. It also has an Audio Overview feature, which can generate a podcast between two voices that deep dives into the content given, which went viral in 2024 when it was first released.
NotebookLM has been particularly well-received for a few reasons. As mentioned above, I think the fact that it allowed people to listen to information being read to them lowered the barrier of entry for them to study. It may have even felt like cheating when they weren’t reading multiple journal articles but instead had a format they associated with a pleasurable and relaxing experience. This also has the effect of being more social media-friendly and thus may have had an easier time gaining views online versus a simple text generator. But I think the biggest contribution to it being what the tech news website Engadget called “the acceptable face of Google AI” is the requirement of the user uploading specific documents as the AI’s only sources. This significantly reduces the likelihood of hallucinations in the output, as it can’t use its own reasoning or outside resources to generate false statistics or insights that aren’t found in what was uploaded. Giving faulty information has been one of the biggest hurdles for AI developers, an issue they have yet to fully solve; however, having a model that is particularly good at avoiding these pitfalls by comparison is going to inevitably score major points with the public.
This is also true for AlphaFold, a program also developed by Google to predict protein structures. The reaction to this model has been even more positive, so much so that the two head developers were awarded the Nobel Prize in Chemistry for their most recent version in 2024. Its most notable scientific contribution was predicting the structures of the virus behind COVID-19, which was later confirmed by comparing AlphaFold’s results to findings from other databases like the Protein Data Bank. It has also contributed to the field of genetics by being a part of the Earth BioGenome Project, which is sequencing the genomes of all the planet’s eukaryotes, that is, organisms that have a nucleus in their cells. This tool sharply decreases both the time and cost of protein structure prediction and has made biologists’ jobs much more efficient.
We see from these examples that the least controversial and most positively received products that have come out of the recent AI boom have identified a specific issue, like the frustration of sifting through mountains of academic papers or the exorbitant investment in time and funding of studying proteins, and sought a reliable way to fix it that doesn’t lead to additional complications. Expanding this methodology for developing AI tools to become the standard may be the best way to satisfy everyone. AI companies should, instead of focusing on a generalized LLM, create models that have a tangible, positive effect on the user’s life. The same concept behind NotebookLM could be applied to other areas.
Imagine a tool that assists with home repair by uploading pictures of the damage and advising on how to fix it. When the prices of anything and everything are at what feels like an all-time-high, having a helper that can give step-by-step instructions for simple repairs could save people thousands instead of hiring someone to come in. It’d also force the assistant to be as accurate as possible lest any hallucination cause even more damage. You can also extend the academic scope to uploading a class syllabus and readings to act as a personal tutor. Students often can’t get the support they need when they’re struggling in school, and a private tutor can also cost thousands of dollars. A personalized tutor tailored to a student’s learning style would not only help close the achievement gap between the most and least advantaged, it’d be a benefit for all students, allowing deeper internalization of the material and eventually leading to a more educated population. It would also reduce a lot of the fear around using AI to cheat, changing the narrative from replacing teachers to augmenting the learning experience. Both these tools democratize access to knowledge that is often too expensive for many average citizens.
This would also be a massive PR boost for them with the public. If people hear that these companies are spending their resources on both creating more practical tools that are meant to make people’s lives easier, while also halting their cutting-edge projects to ensure they don’t cause harm to the environment, the economy, or the social fabric, they become more likely to trust the companies, and once trust is built, the more advanced models can continue to be released. And for those who are interested in AI and its capabilities, having new products in a variety of fields may lead to new innovations and may move society in the direction they desire.
AI companies need to do a better job at understanding what people want and communicating how they are meeting them. When they are seeking to build world-changing technology, it is of the utmost importance that they center how the public perceives them when making decisions. Otherwise, they could be held liable for significant social upheaval if backlash towards them reaches a fever pitch. To avoid such calamity, they must design and develop products that have a very immediate benefit and can integrate into people’s existing daily routines. Once trust is built, society can work towards a proverbial New Eden that truly fulfills the promises tech companies tout.
