Recently Thomas Mirus, Editor-in-Chief of Catholic Culture, invited me (along with many other people in Catholic media) to sign a public pledge he wrote against AI content. After some consideration, I declined to sign it, so I want to explain why.

To be clear, I think the motivations behind the pledge are good, and I’m supportive of the spirit of this particular pledge. Many fine people I highly respect have signed the pledge. Further, we here at Crisis Magazine were one of the first in Catholic media to develop and implement an AI policy for our organization. It states:

We expect all our articles to be fully written by the (human) author, but it is acceptable for writers to use AI for assistance in cleaning up and tightening an article. To this end, we will check every submission with an AI detector and use the results as a significant factor in considering the article for publication.

In other words, we don’t publish AI-generated articles, and we do everything we can to prevent that from happening. Crisis Magazine is an opinion magazine, and we want to publish the opinions of humans, written by humans. While I’m personally supportive of many of the uses of artificial intelligence, I don’t think opinion pieces are something that should ever be offloaded to machines. So, again, I’m in general agreement with those who signed Catholic Culture’s pledge against AI content in Catholic media.

So why didn’t I sign it myself?

First, I admit that I have a natural aversion to signing public pledges, something I inherited from my father, who had a standing policy to refuse all such requests (he was a public figure and so was often asked). My reticence stems from being very particular about how I express my own views. An open pledge like this one naturally expresses the creator’s thoughts, in his words and with his emphases. While I might agree with the overall thrust of what he’s saying, I might disagree with many of the particulars. And in this case, I do disagree with some of the details.

For example, in one instance the pledge declares:

I will verify that any and all research data provided by AI engines is corroborated by non-AI-compromised sources if I make positive use of that information in my published work.

While it’s true that one should be careful about information provided by AI, I think this goes too far. For years we’ve all done research by googling. Depending on the situation, we only sometimes try to verify what we’ve found. If I’m writing a book, especially an academic one, I’m going to check, double-check, and triple-check the sources. But if I’m writing an internet article or preparing for a podcast, I generally am going to trust what I find through Google, unless I have reason not to.

How is that any different from AI doing the research? I know the first response from many will be, “But AI hallucinates!” That’s true, but it’s increasingly less common. In fact, in my own use of AI for research, I haven’t encountered a hallucination in months. So, in general, if it’s a subject I’m already familiar with, I’m not going to verify every single claim made by AI unless I have a good reason. (And I’m not even sure what “non-AI-compromised sources” means.)

Beyond some specifics like this one, I also think the language of the pledge is too broad and sweeping. The pledge includes this statement:

Given the inherent harm of the use of artificial “intelligence” to substitute for human thought, conversation, creativity, and expression, and the potential harms of even its more limited use as a background tool in the production of digital and written material, I have a duty to use AI in a responsible and honest way if I choose to use it at all.

I’m not sure how much I agree with this. I absolutely think AI has the potential to negatively impact cognitive abilities (much like the internet and smartphones have already done). But at the same time, I believe that AI, properly and carefully used, can actually help creativity. It can be a means to stimulate thought. The language of this pledge, however, places AI in a small box, one that I think is too constricting.

The reality is that none of us know exactly the potential benefits and dangers of the use of artificial intelligence. I think it’s great that many people, including the signatories of this pledge, are warning us of the downsides. Most people, including most Catholics, accepted the promised benefits of past technologies like television, the internet, and smartphones without considering the costs. We shouldn’t repeat that mistake with AI (and I detail many of those potential dangers in my upcoming book on artificial intelligence). Yet at the same time we shouldn’t put ourselves in an anti-AI box until we really understand what it can offer (I like to say that I’m not pro-AI as much as I’m anti-anti-AI). In my mind, this pledge seems to do that.

Here at Crisis we will continue to insist that all our content originates with, and is written by, humans, not machines. We’ll be open to valid uses of AI, however, in bringing that content to you, our readers. As Pope Leo wrote, “AI can be a valuable tool,” while also warning that “it calls for a measured and vigilant approach” (Magnifica Humanitas 100). These will be our marching orders here at Crisis Magazine.