AI, Responsibility & Freedom

There are new AI systems today that can do many things that computer systems have never been able to do before. The temptation to avoid effort and let AI systems do our work for us is not overwhelming yet, but it is already there, After all, AI increasingly seems to do a perfectly acceptable job. In some ways this is nothing new: helpful automated systems have existed for a long time, but what is new is the increase in capability in recent years. As these AI systems can do more and more, it is becoming tempting, sometimes very tempting, to let them do our work for us, and pass it off as our own.

This is a temptation that must be resisted, because it is not good for us. I do not say that we should not use AI as a tool to help our work become better. But we need to diligently ensure that the work we do remains our own, and not the work of the AI system itself. If it suggests a turn of phrase or a line of argumentation in our writing, we can consider it as a possibility, but we should also consider our own. If we opt to take its suggestion, it should be a deliberate choice, based entirely on the merits of the suggestion. If, instead, we hand over to the AI system the work itself, perhaps passing a cursory eye over it but really giving it free rein to do what we ourselves should be doing, then there are some very real dangers to us in that.

What “very real dangers” could those be?

The first danger comes out of the fact that our work is our own responsibility. We are accountable for it. If it is wrong, or badly done, that is on us. “The AI did it” is not an acceptable reason for an error, or even a good excuse, it’s an admission of irresponsibility: it says that we handed to an automated system our duty of care. This is true not only for the usual domains of social responsibility, such as our day-to-day duty of care to our co-workers, our clients, our friends, our neighbors and our family, but it is true also for our duty of care before God. For many of us, God has entrusted the good of others to our care. We will have to answer to God about how we carry that out. Even if we entrust our duty to an automated system, we remain responsible for it being done properly.

The second danger is that we are responsible not only for our work, but also to maintain and enhance our ability to do our work. That means we cannot allow our abilities to atrophy, by having someone or something else do our work for us. If we have a responsibility of care, we will need to ensure that we can carry out that responsibility ourselves, because someday we may have to. Relying too heavily on an automated system without exercising and practicing our own ability means that when we need that ability, we may not have it. Yes, it may be convenient to allow someone or something else to do our work for us, but it is not good if it means we become incapable of doing it ourselves.

The third danger, and perhaps the most important, is that we, as human beings, are called to choose what is right. To choose what is right means more than just choosing the right thing when a choice happens to comes up, it also means maintaining our freedom, our ability to choose itself. Here is where the new capabilities of AI-enhanced automated systems are particularly problematic: increasingly, they can choose for us. But if that means that our own ability to make choices becomes impaired, then there is a risk that it will be difficult, or even impossible, to choose what is right when the automated system we rely upon chooses something wrong. But choosing what is right is our duty and responsibility as human beings, a duty and responsibility that is not only necessary for our good and for the good of others, it is something about which God will require from each of us an accounting. He will not accept “the AI did it” as an excuse.

I am not saying that we should avoid the use of AI. In the modern world, that will become less and less possible anyway. Rather, we should ensure that no matter how capable automated AI systems become, we will remain responsible for our actions, we will maintain our ability to carry out our duty, and we will preserve our freedom to do what is right. If we allow our responsibilities to slip away from us, if we allow our abilities to atrophy, and we relinquish our freedom, we will not be able to stand before God on the last day and give an adequate account of our lives. Let us be aware of the danger now, and take steps today to ensure that no matter how capable AI systems become, we will always remain responsible, able and free, to do what we have been called to do.