AI and Human Responsibility

We must resist the machine and its attempts to dehumanise us:

The worrying tendency of humans to play down human responsibility or to allow others to make choices for them has of course always been with us. This includes allowing various “experts” to dictate to us what is right and wrong, true and false. It is often easier to defer to others.

But in our super-technological age, this is getting even more pronounced – and worse. Allowing machines – no matter how ‘smart’ these machines might be – to make choices for us is all the rage. And as a result, the abdication of human responsibility and accountability is ramping up.

The world of AI is a primary example of this. We are letting AI do everything for us, from writing a book, driving our cars, engaging in artwork, or even preparing a sermon. As we grant AI more and more authority and free rein in determining how we live and what choices we make, we really are heading into a brave new world. We really are allowing for our own dehumanisation.

Too many folks – Christians included – just look at all the supposed benefits of AI and the new technologies. And to be sure, there are benefits. But there are also real downsides, and not thinking carefully and critically about where the AI revolution is taking us is a recipe for disaster.

One need not be a writer of ominous sci fi novels to speak this way. Increasingly, serious social thinkers, philosophers, theologians, scientists and others are sounding the alarm. And well they should. I have long been alerting my readers to important books on this.

In a recent piece I featured three new volumes that are urging us to think hard and long about the AI journey that we are all on. One of the books I discussed was Servants of the Machine: AI and the Abdication of Human Responsibility by Simon Christian Trepp (Quaternity Editions, 2026). See some quotes from that book here: https://billmuehlenberg.com/2026/07/16/ai-god-and-the-christian/

I want to share some more quotes from this volume. They all come from his concluding chapters. He speaks of the dangers of the “liquefying of responsibility”:

If we get too accustomed to not feel responsible for our choices, we might carry that habit into areas where human judgment is essential. Today it might be how you choose a playlist. Tomorrow it might be how a judge sentences a defendant, with a risk assessment AI whispering in their ear. The pattern is the same, only the stakes differ….

 

[T]he pattern is psychologically seductive. It offers relief from the burden of judgment. When we over-rely on complex tools, we gradually lose competence and autonomy. We become dependent not just on the tool’s function, but on its judgment. Our own skills, whether map-reading, critical reading and thinking, or moral reasoning, quietly atrophy from disuse.…

 

Today’s abdication is quiet and incremental. We are not rallying behind a charismatic dictator. We are sliding behind a constellation of apps and systems that make our choices for us. The result may not be a boot-on-the-neck tyranny, but it could manifest as a soft despotism of compliance. A society in which people largely do what the systems recommend because it’s easy and largely effective, all the while losing the habit of challenging or diverging from those recommendations. (pp. 83-84)

He goes on to say this:

The death of the moral actor is ultimately the death of the idea that each of us has a duty to judge right from wrong and act accordingly. In a fully optimised, algorithmically managed world, the very occasions for moral judgment shrink. If a self-driving car must choose between two accident outcomes, and the decision has been pre-loaded by programming, no human driver makes an ethical choice in the moment. It’s decided at design time, perhaps without moral fanfare, perhaps by engineers considering probabilities and liabilities more than moral philosophy. If down the line someone questions that choice, say the car sacrificed one life for another, the company can say it followed industry standards or regulatory guidelines, again dispersing responsibility. The actual moment of choice was so embedded in technical processes that it ceased to register as a moral moment for anyone. (p. 91)

Trepp says the way forward is not to abandon or reject all the new technologies, but to put

technology back in its place, as a servant of human ends, not a surrogate for human responsibility. It means remembering that AI is infrastructure, not an independent agent. We should not bow to our own tools or fear them, as gods. We should manage them as tools, with humans clearly accountable for their use. It also means preserving spaces where human judgment can flourish without being overruled by optimisation. Spaces of education, deliberation, and creativity where inefficiency and even error are tolerated because they are essential to growth and wisdom.

 

Comfort can be a quiet road to servitude. The time has come to exchange a bit of that comfort for a more demanding freedom. (pp. 97-98)

He again restates and reemphasises the importance of moral agency:

Judgement is like a muscle. If you never use it unassisted, it weakens. One of the great errors of the modern technological imagination is to treat human judgement as a defect to be engineered away. But judgement is not a bug. It is the seat of moral life. It is where a person wrestles with values. It is where meaning is made.

 

A person who never makes difficult choices has an easy life, but a thin one. They may be safe. They may also be hollow. (p. 107)

And again:

Freedom is not neat. Agency is not optimal. Responsibility is heavy. It demands time, doubt, and sometimes conflict. Abdication is easy. Agency is hard. It is far more comfortable to say, “I was just following the algorithm” than to say, “I chose this, and I stand behind it.”

 

The modern world is built to make that discomfort feel unnecessary. It offers us endless ways to avoid being the one who decides.

 

But the discomfort is the price of a morally meaningful life. A decision that could not have been otherwise, because the system ordained it, carries no moral weight. A decision in which a person struggled, weighed values, and took a stance becomes part of their moral story. It may be wrong. It is still human. It can be owned. It can be repaired. It can be answered for.

 

A machine cannot do that. A machine cannot feel remorse. A machine cannot atone. A machine cannot be reconciled with. A machine cannot stand in front of another person and say, without abstraction, “I did this”.

 

Only a person can. (p. 109)

The book’s closing paragraphs are these:

The question is no longer what the machines will do. The question is whether we will continue to behave as if we require an unquestionable voice. No system will take our agency from us without our cooperation. No algorithm can absolve us unless we wish to be absolved. The oracle only speaks because we keep listening.

 

If there is a task ahead, it is not to defeat technology, nor to sanctify it. It is to remain human in its presence.

 

That means deciding. It means answering for what we decide. It means refusing the relief of abdication, even when it is offered politely and efficiently.

 

We do not need new gods. We need the courage to live without them. The rest will follow from that. (p. 113)

Wise words indeed. We need to tame the machine gods before they tame – or eliminate – us.

[1258 words]