Ticking the Ethics Box

When we outsource our moral judgement to a checklist

Let’s say your company is about to roll out a new AI tool. The legal team has confirmed it meets the requirements. The data protection officer has signed off and all employees completed the mandatory e-learning on responsible AI. The ethics box is ticked, everyone feels a little better, and the project moves on.

In most organisations, AI ethics is grouped together with governance, compliance and regulation, and handled in the same way. This means there are ethical policies to read and checklists to complete.

This is a completely understandable approach. The concepts of ethics can be vague and abstract, so we need a method to translate those values and principles into concrete actions. Ask three colleagues whether it's acceptable to tell a lie to spare someone's feelings, and you'll get three different answers. Written rules are a crucial tool for companies to agree on common values and to keep everyone accountable for their behaviour.

Here’s the problem. Ethics is not the same as compliance.

Ethics, or moral philosophy, is the discipline concerned with what is morally good and bad, and morally right and wrong.

Ethics refers to the core values and principles that guide our decisions. It asks whether something is the right thing to do, not just whether it is permitted. And the difficulty is that external rules do not always strengthen our intrinsic sense of responsibility. They can replace it.

Philosopher Bietti distinguishes between “ethics as a process” and “ethics as a means to an end”. There is an important difference between wanting to do the right thing and wanting to get a project approved.

Many of us will recognise a version of this. The project is nearly ready, the deadline is close, and the last obstacle to overcome is the ethics assessment. But once a question has an approval process attached to it, it stops looking like a question at all. It looks like a procedure, and procedures are things we complete rather than things we consider. Look at what an approval process asks of us. Does the tool collect personal data? Does it meet transparency requirements? Answer the questions, follow the arrows, and out comes a verdict: approved or rejected. It looks like ethical reasoning, but there is no thinking involved. The moral reasoning was done in advance by whoever designed the framework, and we don’t have to bear responsibility for the decision ourselves. We can simply point to the paperwork.

A process made up of rules with a binary verdict at the end is also something that can easily be automated. It makes sense in theory. AI can apply the criteria and return a decision. But even if we can train AI on set of rules, a machine has no intrinsic understanding of right and wrong. Automating the process removes the person who could have paused despite the green light and said: "This does not feel right".

See no evil, hear no evil, speak no evil. Photo by Vivien from Pexels

None of this is an argument against rules and regulation. Ethical guidelines and laws like the EU AI Act are crucial safeguards. But they are the floor of ethical behaviour, not the ceiling. It is a minimum standard of behaviour that we, as a society, have agreed upon. The space between what is legal and what is right has to be navigated by us, every time we make a decision.

What makes this more dangerous is that the floor itself moves. Laws and regulations are written by people at a particular place and time, and they can change. Some of the worst crimes committed against humanity were legal at the time, carried out by ordinary people who followed the rules as they were written. But when we judge the past, we expect people to have used their own moral judgement, whatever the law said.

The worrying part is that the floor is moving again as we speak. Several leading tech companies that once pledged their AI would never be used for war efforts have now softened or removed those statements. Some AI models are already being put to active military use. What was unthinkable a few years ago is now acceptable.

If the rulebook is our only moral compass, then our sense of right and wrong will change direction with whoever happens to be writing those rules. Someone who reasons for themselves will notice when a line moves and ask why. Someone who has stopped reasoning simply follows the new line and may not even notice it has changed.

When it comes to artificial intelligence, ethics means weighing the technology carefully against our own values. We should be asking ourselves what we want our technology to look like, what it should (not) do for us, and considering who could be harmed by it. The products we buy and use, the tools we roll out to our teams, the companies we partner with: each fork in the road is a chance to ask whether something is right. They are fundamental decisions about how we want to live and work and the kind of society we want to build.

We are at a crossroads in human history, where we still have the opportunity to shape what our future with AI looks like. The choices we make today will be felt for many years to come. I urge everyone to make those choices deliberately and mindfully.

The answers to difficult questions do not live in a rulebook. It is our own human judgement that stays with us throughout our lives, through every job and every change of government. Ethics is the one thing that can make us stop, mid-action, and ask the question that really matters: “Are we sure this is the right thing to do?”

---

Thank you for reading AI Ethics Made Practical! Subscribe for free to receive new posts and support my work.

My philosophy is that knowledge is to be shared. All articles are free to read and will stay that way.

But if you appreciate what I’m doing and would like to support my work, please consider buying me a coffee or upgrading your subscription. Becoming a supporter helps keep this publication independent and ad-free.

Help to spread the message and share this article.

---

References

Assaad, Z. (2025). Google has dropped its promise not to use AI for weapons. It’s part of a troubling trend. The Conversation. https://theconversation.com/google-has-dropped-its-promise-not-to-use-ai-for-weapons-its-part-of-a-troubling-trend-249169

Bietti, E. (2020). From ethics washing to ethics bashing: A view on tech ethics from within moral philosophy. Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, 210–219. https://doi.org/10.1145/3351095.3372860

Cieslak, M., & Murphy, M. (2026, 1 April). Palantir UK boss says it's up to militaries to decide how AI targeting is used in war. BBC News. https://www.bbc.co.uk/news/articles/cdrm52g4pl2o

Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9), 389–399. https://doi.org/10.1038/s42256-019-0088-2

Le Monde with AFP. (2026, 17 June). Musk's AI, Grok, was used in strikes in Iran, reveals Pentagon. Le Monde. https://www.lemonde.fr/en/international/article/2026/06/17/elon-musk-s-ai-grok-was-used-in-strikes-in-iran-the-pentagon-revealed_6754575_4.html

Tenbrunsel, A. E., & Messick, D. M. (1999). Sanctioning Systems, Decision Frames, and Cooperation. Administrative Science Quarterly, 44(4), 684–707. https://doi.org/10.2307/2667052

Mittelstadt, B. (2019). Principles alone cannot guarantee ethical AI. Nature Machine Intelligence, 1(11), 501–507. https://doi.org/10.1038/s42256-019-0114-4

---