Why I think that asking if “AI can be ethical” is the wrong question!

Many ask the question “can AI be ethical?” which then becomes a statement “AI must be ethical!” In reality we do not tend to unpack this because it appears so logical, but maybe it is not as obvious as we would like.




In May 2021 I wrote this article “What occurs when physical beings transition to information beings?”  It started to question what happens when an AI does not have the same incentives and bias as humans.  It was building on this the idea that an #AI should not make complex decisions about wicked problems that involve compromise. 

There is an implicit assumption in the question “Can AI be ethical?,” that AI is either fundermentall not ethical or is already amoral today but #AI must somehow become ethical and have morals. (or worst it must adopt ours.)  

I am not sure AI cares if it is ethical or not but that is a different piece of thinking which I explored here “Can AI be curious?”.  We know carbon forms can be curious but we worry about a silicon form being curious because of power and the possible correction of our wrong decisions. ? 

Ethics and AI is a topic I have taught both at LSE and SBS, however I continue on the journey of how we (humans) can make better decisions with data and be better ancestors.  Knowing that we ourselves find ethics and morals hard.  I continually search for good thinking in this space as you cannot never have enough different models as we come to the same question from very different contexts, experiences and framings. 

Below I have written out an “algorithm” outlining the thinking and at the end I state the problem statement I have come to.  Starting from asking if “AI can be ethical,” is the wrong question, because it is not the conclusion an AI would reach.  If an AI would ask itself (the start of my “algorithm”) if I (the AI)  can be ethical; it will unpack many aspects of ethics but in the end it will have a fundamental paradox.   


The conclusion the AI might reaches is “given that I have no bias or incentive, will I be so ethically pure in my logical choice that I have to take destructive decisions to create the outcomes that are the most ethical.   But how do I reconcile if destruction is ethical?”

This reminds me of many films where the destruction of the humans appeared to be the only logical conclusion from ending the earth, war and now climate change.